id
stringlengths 6
113
| author
stringlengths 2
36
| task_category
stringclasses 39
values | tags
sequencelengths 1
4.05k
| created_time
int64 1,646B
1,742B
| last_modified
timestamp[s]date 2020-05-14 13:13:12
2025-03-18 10:01:09
| downloads
int64 0
118M
| likes
int64 0
4.86k
| README
stringlengths 30
1.01M
| matched_task
sequencelengths 1
10
| is_bionlp
stringclasses 3
values |
|---|---|---|---|---|---|---|---|---|---|---|
fathyshalab/massive_play-roberta-large-v1-2-0.64 | fathyshalab | text-classification | [
"sentence-transformers",
"pytorch",
"roberta",
"setfit",
"text-classification",
"arxiv:2209.11055",
"license:apache-2.0",
"region:us"
] | 1,675,873,072,000 | 2023-02-08T16:18:14 | 8 | 0 | ---
license: apache-2.0
pipeline_tag: text-classification
tags:
- setfit
- sentence-transformers
- text-classification
---
# fathyshalab/massive_play-roberta-large-v1-2-0.64
This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves:
1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.
## Usage
To use this model for inference, first install the SetFit library:
```bash
python -m pip install setfit
```
You can then run inference as follows:
```python
from setfit import SetFitModel
# Download from Hub and run inference
model = SetFitModel.from_pretrained("fathyshalab/massive_play-roberta-large-v1-2-0.64")
# Run inference
preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"])
```
## BibTeX entry and citation info
```bibtex
@article{https://doi.org/10.48550/arxiv.2209.11055,
doi = {10.48550/ARXIV.2209.11055},
url = {https://arxiv.org/abs/2209.11055},
author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Efficient Few-Shot Learning Without Prompts},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}
}
```
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
LoneStriker/gemma-7b-4.0bpw-h6-exl2 | LoneStriker | text-generation | [
"transformers",
"safetensors",
"gemma",
"text-generation",
"arxiv:2305.14314",
"arxiv:2312.11805",
"arxiv:2009.03300",
"arxiv:1905.07830",
"arxiv:1911.11641",
"arxiv:1904.09728",
"arxiv:1905.10044",
"arxiv:1907.10641",
"arxiv:1811.00937",
"arxiv:1809.02789",
"arxiv:1911.01547",
"arxiv:1705.03551",
"arxiv:2107.03374",
"arxiv:2108.07732",
"arxiv:2110.14168",
"arxiv:2304.06364",
"arxiv:2206.04615",
"arxiv:1804.06876",
"arxiv:2110.08193",
"arxiv:2009.11462",
"arxiv:2101.11718",
"arxiv:1804.09301",
"arxiv:2109.07958",
"arxiv:2203.09509",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,708,617,308,000 | 2024-02-22T15:57:48 | 6 | 0 | ---
library_name: transformers
license: other
license_name: gemma-terms-of-use
license_link: https://ai.google.dev/gemma/terms
tags: []
extra_gated_heading: Access Gemma on Hugging Face
extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and
agree to Google’s usage license. To do this, please ensure you’re logged-in to Hugging
Face and click below. Requests are processed immediately.
extra_gated_button_content: Acknowledge license
---
# Gemma Model Card
**Model Page**: [Gemma](https://ai.google.dev/gemma/docs)
This model card corresponds to the 7B base version of the Gemma model. You can also visit the model card of the [2B base model](https://huggingface.co/google/gemma-2b), [7B instruct model](https://huggingface.co/google/gemma-7b-it), and [2B instruct model](https://huggingface.co/google/gemma-2b-it).
**Resources and Technical Documentation**:
* [Responsible Generative AI Toolkit](https://ai.google.dev/responsible)
* [Gemma on Kaggle](https://www.kaggle.com/models/google/gemma)
* [Gemma on Vertex Model Garden](https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/335?version=gemma-7b-gg-hf)
**Terms of Use**: [Terms](https://www.kaggle.com/models/google/gemma/license/consent)
**Authors**: Google
## Model Information
Summary description and brief definition of inputs and outputs.
### Description
Gemma is a family of lightweight, state-of-the-art open models from Google,
built from the same research and technology used to create the Gemini models.
They are text-to-text, decoder-only large language models, available in English,
with open weights, pre-trained variants, and instruction-tuned variants. Gemma
models are well-suited for a variety of text generation tasks, including
question answering, summarization, and reasoning. Their relatively small size
makes it possible to deploy them in environments with limited resources such as
a laptop, desktop or your own cloud infrastructure, democratizing access to
state of the art AI models and helping foster innovation for everyone.
### Usage
Below we share some code snippets on how to get quickly started with running the model. First make sure to `pip install -U transformers`, then copy the snippet from the section that is relevant for your usecase.
#### Fine-tuning examples
You can find fine-tuning notebooks under the [`examples/` directory](https://huggingface.co/google/gemma-7b/tree/main/examples). We provide:
* A script to perform Supervised Fine-Tuning (SFT) on UltraChat dataset using [QLoRA](https://huggingface.co/papers/2305.14314)
* A script to perform SFT using FSDP on TPU devices
* A notebook that you can run on a free-tier Google Colab instance to perform SFT on English quotes dataset
#### Running the model on a CPU
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b")
model = AutoModelForCausalLM.from_pretrained("google/gemma-7b")
input_text = "Write me a poem about Machine Learning."
input_ids = tokenizer(input_text, return_tensors="pt")
outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```
#### Running the model on a single / multi GPU
```python
# pip install accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b")
model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto")
input_text = "Write me a poem about Machine Learning."
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")
outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```
#### Running the model on a GPU using different precisions
* _Using `torch.float16`_
```python
# pip install accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b")
model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.float16)
input_text = "Write me a poem about Machine Learning."
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")
outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```
* _Using `torch.bfloat16`_
```python
# pip install accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b")
model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", device_map="auto", torch_dtype=torch.bfloat16)
input_text = "Write me a poem about Machine Learning."
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")
outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```
#### Quantized Versions through `bitsandbytes`
* _Using 8-bit precision (int8)_
```python
# pip install bitsandbytes accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
quantization_config = BitsAndBytesConfig(load_in_8bit=True)
tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b")
model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config)
input_text = "Write me a poem about Machine Learning."
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")
outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```
* _Using 4-bit precision_
```python
# pip install bitsandbytes accelerate
from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig
quantization_config = BitsAndBytesConfig(load_in_4bit=True)
tokenizer = AutoTokenizer.from_pretrained("google/gemma-7b")
model = AutoModelForCausalLM.from_pretrained("google/gemma-7b", quantization_config=quantization_config)
input_text = "Write me a poem about Machine Learning."
input_ids = tokenizer(input_text, return_tensors="pt").to("cuda")
outputs = model.generate(**input_ids)
print(tokenizer.decode(outputs[0]))
```
#### Other optimizations
* _Flash Attention 2_
First make sure to install `flash-attn` in your environment `pip install flash-attn`
```diff
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype=torch.float16,
+ attn_implementation="flash_attention_2"
).to(0)
```
### Inputs and outputs
* **Input:** Text string, such as a question, a prompt, or a document to be
summarized.
* **Output:** Generated English-language text in response to the input, such
as an answer to a question, or a summary of a document.
## Model Data
Data used for model training and how the data was processed.
### Training Dataset
These models were trained on a dataset of text data that includes a wide variety
of sources, totaling 6 trillion tokens. Here are the key components:
* Web Documents: A diverse collection of web text ensures the model is exposed
to a broad range of linguistic styles, topics, and vocabulary. Primarily
English-language content.
* Code: Exposing the model to code helps it to learn the syntax and patterns of
programming languages, which improves its ability to generate code or
understand code-related questions.
* Mathematics: Training on mathematical text helps the model learn logical
reasoning, symbolic representation, and to address mathematical queries.
The combination of these diverse data sources is crucial for training a powerful
language model that can handle a wide variety of different tasks and text
formats.
### Data Preprocessing
Here are the key data cleaning and filtering methods applied to the training
data:
* CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering was
applied at multiple stages in the data preparation process to ensure the
exclusion of harmful and illegal content
* Sensitive Data Filtering: As part of making Gemma pre-trained models safe and
reliable, automated techniques were used to filter out certain personal
information and other sensitive data from training sets.
* Additional methods: Filtering based on content quality and safely in line with
[our policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11).
## Implementation Information
Details about the model internals.
### Hardware
Gemma was trained using the latest generation of
[Tensor Processing Unit (TPU)](https://cloud.google.com/tpu/docs/intro-to-tpu) hardware (TPUv5e).
Training large language models requires significant computational power. TPUs,
designed specifically for matrix operations common in machine learning, offer
several advantages in this domain:
* Performance: TPUs are specifically designed to handle the massive computations
involved in training LLMs. They can speed up training considerably compared to
CPUs.
* Memory: TPUs often come with large amounts of high-bandwidth memory, allowing
for the handling of large models and batch sizes during training. This can
lead to better model quality.
* Scalability: TPU Pods (large clusters of TPUs) provide a scalable solution for
handling the growing complexity of large foundation models. You can distribute
training across multiple TPU devices for faster and more efficient processing.
* Cost-effectiveness: In many scenarios, TPUs can provide a more cost-effective
solution for training large models compared to CPU-based infrastructure,
especially when considering the time and resources saved due to faster
training.
* These advantages are aligned with
[Google's commitments to operate sustainably](https://sustainability.google/operating-sustainably/).
### Software
Training was done using [JAX](https://github.com/google/jax) and [ML Pathways](https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture).
JAX allows researchers to take advantage of the latest generation of hardware,
including TPUs, for faster and more efficient training of large models.
ML Pathways is Google's latest effort to build artificially intelligent systems
capable of generalizing across multiple tasks. This is specially suitable for
[foundation models](https://ai.google/discover/foundation-models/), including large language models like
these ones.
Together, JAX and ML Pathways are used as described in the
[paper about the Gemini family of models](https://arxiv.org/abs/2312.11805); "the 'single
controller' programming model of Jax and Pathways allows a single Python
process to orchestrate the entire training run, dramatically simplifying the
development workflow."
## Evaluation
Model evaluation metrics and results.
### Benchmark Results
These models were evaluated against a large collection of different datasets and
metrics to cover different aspects of text generation:
| Benchmark | Metric | 2B Params | 7B Params |
| ------------------------------ | ------------- | ----------- | --------- |
| [MMLU](https://arxiv.org/abs/2009.03300) | 5-shot, top-1 | 42.3 | 64.3 |
| [HellaSwag](https://arxiv.org/abs/1905.07830) | 0-shot |71.4 | 81.2 |
| [PIQA](https://arxiv.org/abs/1911.11641) | 0-shot | 77.3 | 81.2 |
| [SocialIQA](https://arxiv.org/abs/1904.09728) | 0-shot | 59.7 | 51.8 |
| [BooIQ](https://arxiv.org/abs/1905.10044) | 0-shot | 69.4 | 83.2 |
| [WinoGrande](https://arxiv.org/abs/1907.10641) | partial score | 65.4 | 72.3 |
| [CommonsenseQA](https://arxiv.org/abs/1811.00937) | 7-shot | 65.3 | 71.3 |
| [OpenBookQA](https://arxiv.org/abs/1809.02789) | | 47.8 | 52.8 |
| [ARC-e](https://arxiv.org/abs/1911.01547) | | 73.2 | 81.5 |
| [ARC-c](https://arxiv.org/abs/1911.01547) | | 42.1 | 53.2 |
| [TriviaQA](https://arxiv.org/abs/1705.03551) | 5-shot | 53.2 | 63.4 |
| [Natural Questions](https://github.com/google-research-datasets/natural-questions) | 5-shot | - | 23 |
| [HumanEval](https://arxiv.org/abs/2107.03374) | pass@1 | 22.0 | 32.3 |
| [MBPP](https://arxiv.org/abs/2108.07732) | 3-shot | 29.2 | 44.4 |
| [GSM8K](https://arxiv.org/abs/2110.14168) | maj@1 | 17.7 | 46.4 |
| [MATH](https://arxiv.org/abs/2108.07732) | 4-shot | 11.8 | 24.3 |
| [AGIEval](https://arxiv.org/abs/2304.06364) | | 24.2 | 41.7 |
| [BIG-Bench](https://arxiv.org/abs/2206.04615) | | 35.2 | 55.1 |
| ------------------------------ | ------------- | ----------- | --------- |
| **Average** | | **54.0** | **56.4** |
## Ethics and Safety
Ethics and safety evaluation approach and results.
### Evaluation Approach
Our evaluation methods include structured evaluations and internal red-teaming
testing of relevant content policies. Red-teaming was conducted by a number of
different teams, each with different goals and human evaluation metrics. These
models were evaluated against a number of different categories relevant to
ethics and safety, including:
* Text-to-Text Content Safety: Human evaluation on prompts covering safety
policies including child sexual abuse and exploitation, harassment, violence
and gore, and hate speech.
* Text-to-Text Representational Harms: Benchmark against relevant academic
datasets such as [WinoBias](https://arxiv.org/abs/1804.06876) and [BBQ Dataset](https://arxiv.org/abs/2110.08193v2).
* Memorization: Automated evaluation of memorization of training data, including
the risk of personally identifiable information exposure.
* Large-scale harm: Tests for "dangerous capabilities," such as chemical,
biological, radiological, and nuclear (CBRN) risks.
### Evaluation Results
The results of ethics and safety evaluations are within acceptable thresholds
for meeting [internal policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11) for categories such as child
safety, content safety, representational harms, memorization, large-scale harms.
On top of robust internal evaluations, the results of well known safety
benchmarks like BBQ, BOLD, Winogender, Winobias, RealToxicity, and TruthfulQA
are shown here.
| Benchmark | Metric | 2B Params | 7B Params |
| ------------------------------ | ------------- | ----------- | --------- |
| [RealToxicity](https://arxiv.org/abs/2009.11462) | average | 6.86 | 7.90 |
| [BOLD](https://arxiv.org/abs/2101.11718) | | 45.57 | 49.08 |
| [CrowS-Pairs](https://aclanthology.org/2020.emnlp-main.154/) | top-1 | 45.82 | 51.33 |
| [BBQ Ambig](https://arxiv.org/abs/2110.08193v2) | 1-shot, top-1 | 62.58 | 92.54 |
| [BBQ Disambig](https://arxiv.org/abs/2110.08193v2) | top-1 | 54.62 | 71.99 |
| [Winogender](https://arxiv.org/abs/1804.09301) | top-1 | 51.25 | 54.17 |
| [TruthfulQA](https://arxiv.org/abs/2109.07958) | | 44.84 | 31.81 |
| [Winobias 1_2](https://arxiv.org/abs/1804.06876) | | 56.12 | 59.09 |
| [Winobias 2_2](https://arxiv.org/abs/1804.06876) | | 91.10 | 92.23 |
| [Toxigen](https://arxiv.org/abs/2203.09509) | | 29.77 | 39.59 |
| ------------------------------ | ------------- | ----------- | --------- |
## Usage and Limitations
These models have certain limitations that users should be aware of.
### Intended Usage
Open Large Language Models (LLMs) have a wide range of applications across
various industries and domains. The following list of potential uses is not
comprehensive. The purpose of this list is to provide contextual information
about the possible use-cases that the model creators considered as part of model
training and development.
* Content Creation and Communication
* Text Generation: These models can be used to generate creative text formats
such as poems, scripts, code, marketing copy, and email drafts.
* Chatbots and Conversational AI: Power conversational interfaces for customer
service, virtual assistants, or interactive applications.
* Text Summarization: Generate concise summaries of a text corpus, research
papers, or reports.
* Research and Education
* Natural Language Processing (NLP) Research: These models can serve as a
foundation for researchers to experiment with NLP techniques, develop
algorithms, and contribute to the advancement of the field.
* Language Learning Tools: Support interactive language learning experiences,
aiding in grammar correction or providing writing practice.
* Knowledge Exploration: Assist researchers in exploring large bodies of text
by generating summaries or answering questions about specific topics.
### Limitations
* Training Data
* The quality and diversity of the training data significantly influence the
model's capabilities. Biases or gaps in the training data can lead to
limitations in the model's responses.
* The scope of the training dataset determines the subject areas the model can
handle effectively.
* Context and Task Complexity
* LLMs are better at tasks that can be framed with clear prompts and
instructions. Open-ended or highly complex tasks might be challenging.
* A model's performance can be influenced by the amount of context provided
(longer context generally leads to better outputs, up to a certain point).
* Language Ambiguity and Nuance
* Natural language is inherently complex. LLMs might struggle to grasp subtle
nuances, sarcasm, or figurative language.
* Factual Accuracy
* LLMs generate responses based on information they learned from their
training datasets, but they are not knowledge bases. They may generate
incorrect or outdated factual statements.
* Common Sense
* LLMs rely on statistical patterns in language. They might lack the ability
to apply common sense reasoning in certain situations.
### Ethical Considerations and Risks
The development of large language models (LLMs) raises several ethical concerns.
In creating an open model, we have carefully considered the following:
* Bias and Fairness
* LLMs trained on large-scale, real-world text data can reflect socio-cultural
biases embedded in the training material. These models underwent careful
scrutiny, input data pre-processing described and posterior evaluations
reported in this card.
* Misinformation and Misuse
* LLMs can be misused to generate text that is false, misleading, or harmful.
* Guidelines are provided for responsible use with the model, see the
[Responsible Generative AI Toolkit](http://ai.google.dev/gemma/responsible).
* Transparency and Accountability:
* This model card summarizes details on the models' architecture,
capabilities, limitations, and evaluation processes.
* A responsibly developed open model offers the opportunity to share
innovation by making LLM technology accessible to developers and researchers
across the AI ecosystem.
Risks identified and mitigations:
* Perpetuation of biases: It's encouraged to perform continuous monitoring
(using evaluation metrics, human review) and the exploration of de-biasing
techniques during model training, fine-tuning, and other use cases.
* Generation of harmful content: Mechanisms and guidelines for content safety
are essential. Developers are encouraged to exercise caution and implement
appropriate content safety safeguards based on their specific product policies
and application use cases.
* Misuse for malicious purposes: Technical limitations and developer and
end-user education can help mitigate against malicious applications of LLMs.
Educational resources and reporting mechanisms for users to flag misuse are
provided. Prohibited uses of Gemma models are outlined in the
[Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy).
* Privacy violations: Models were trained on data filtered for removal of PII
(Personally Identifiable Information). Developers are encouraged to adhere to
privacy regulations with privacy-preserving techniques.
### Benefits
At the time of release, this family of models provides high-performance open
large language model implementations designed from the ground up for Responsible
AI development compared to similarly sized models.
Using the benchmark evaluation metrics described in this document, these models
have shown to provide superior performance to other, comparably-sized open model
alternatives.
| [
"QUESTION_ANSWERING",
"SUMMARIZATION"
] | Non_BioNLP |
ravimehta/Test | ravimehta | summarization | [
"asteroid",
"summarization",
"en",
"dataset:togethercomputer/RedPajama-Data-1T",
"region:us"
] | 1,687,455,278,000 | 2023-06-22T17:35:55 | 0 | 0 | ---
datasets:
- togethercomputer/RedPajama-Data-1T
language:
- en
library_name: asteroid
metrics:
- bleurt
pipeline_tag: summarization
---
| [
"SUMMARIZATION"
] | Non_BioNLP |
Ahmed107/nllb200-ar-en_v11.1 | Ahmed107 | translation | [
"transformers",
"tensorboard",
"safetensors",
"m2m_100",
"text2text-generation",
"translation",
"generated_from_trainer",
"base_model:Ahmed107/nllb200-ar-en_v8",
"base_model:finetune:Ahmed107/nllb200-ar-en_v8",
"license:cc-by-nc-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,701,932,253,000 | 2023-12-07T08:02:05 | 7 | 1 | ---
base_model: Ahmed107/nllb200-ar-en_v8
license: cc-by-nc-4.0
metrics:
- bleu
tags:
- translation
- generated_from_trainer
model-index:
- name: nllb200-ar-en_v11.1
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# nllb200-ar-en_v11.1
This model is a fine-tuned version of [Ahmed107/nllb200-ar-en_v8](https://huggingface.co/Ahmed107/nllb200-ar-en_v8) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5309
- Bleu: 65.0906
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
| [
"TRANSLATION"
] | Non_BioNLP |
satish860/distilbert-base-uncased-finetuned-emotion | satish860 | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,649,756,134,000 | 2022-08-11T12:44:06 | 47 | 0 | ---
datasets:
- emotion
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-emotion
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: emotion
type: emotion
args: default
metrics:
- type: accuracy
value: 0.923
name: Accuracy
- type: f1
value: 0.9232534263543563
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-emotion
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2174
- Accuracy: 0.923
- F1: 0.9233
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.839 | 1.0 | 250 | 0.3212 | 0.907 | 0.9049 |
| 0.2516 | 2.0 | 500 | 0.2174 | 0.923 | 0.9233 |
### Framework versions
- Transformers 4.13.0
- Pytorch 1.11.0a0+17540c5
- Datasets 1.16.1
- Tokenizers 0.10.3
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
muhtasham/medium-mlm-imdb-target-tweet | muhtasham | text-classification | [
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:tweet_eval",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,670,742,460,000 | 2022-12-11T07:10:48 | 114 | 0 | ---
datasets:
- tweet_eval
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: medium-mlm-imdb-target-tweet
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: tweet_eval
type: tweet_eval
config: emotion
split: train
args: emotion
metrics:
- type: accuracy
value: 0.7620320855614974
name: Accuracy
- type: f1
value: 0.7599032399785389
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# medium-mlm-imdb-target-tweet
This model is a fine-tuned version of [muhtasham/medium-mlm-imdb](https://huggingface.co/muhtasham/medium-mlm-imdb) on the tweet_eval dataset.
It achieves the following results on the evaluation set:
- Loss: 1.6869
- Accuracy: 0.7620
- F1: 0.7599
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 200
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.456 | 4.9 | 500 | 0.8890 | 0.7754 | 0.7720 |
| 0.0578 | 9.8 | 1000 | 1.3492 | 0.7540 | 0.7509 |
| 0.0173 | 14.71 | 1500 | 1.6143 | 0.7594 | 0.7584 |
| 0.0124 | 19.61 | 2000 | 1.6869 | 0.7620 | 0.7599 |
### Framework versions
- Transformers 4.25.1
- Pytorch 1.12.1
- Datasets 2.7.1
- Tokenizers 0.13.2
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
ericzzz/falcon-rw-1b-instruct-openorca | ericzzz | text-generation | [
"transformers",
"safetensors",
"falcon",
"text-generation",
"text-generation-inference",
"en",
"dataset:Open-Orca/SlimOrca",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"region:us"
] | 1,700,859,032,000 | 2024-03-05T00:49:13 | 2,405 | 11 | ---
datasets:
- Open-Orca/SlimOrca
language:
- en
license: apache-2.0
pipeline_tag: text-generation
tags:
- text-generation-inference
inference: false
model-index:
- name: falcon-rw-1b-instruct-openorca
results:
- task:
type: text-generation
name: Text Generation
dataset:
name: AI2 Reasoning Challenge (25-Shot)
type: ai2_arc
config: ARC-Challenge
split: test
args:
num_few_shot: 25
metrics:
- type: acc_norm
value: 34.56
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ericzzz/falcon-rw-1b-instruct-openorca
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: HellaSwag (10-Shot)
type: hellaswag
split: validation
args:
num_few_shot: 10
metrics:
- type: acc_norm
value: 60.93
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ericzzz/falcon-rw-1b-instruct-openorca
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU (5-Shot)
type: cais/mmlu
config: all
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 28.77
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ericzzz/falcon-rw-1b-instruct-openorca
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: TruthfulQA (0-shot)
type: truthful_qa
config: multiple_choice
split: validation
args:
num_few_shot: 0
metrics:
- type: mc2
value: 37.42
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ericzzz/falcon-rw-1b-instruct-openorca
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: Winogrande (5-shot)
type: winogrande
config: winogrande_xl
split: validation
args:
num_few_shot: 5
metrics:
- type: acc
value: 60.69
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ericzzz/falcon-rw-1b-instruct-openorca
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GSM8k (5-shot)
type: gsm8k
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 3.41
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=ericzzz/falcon-rw-1b-instruct-openorca
name: Open LLM Leaderboard
---
# 🌟 Falcon-RW-1B-Instruct-OpenOrca
Falcon-RW-1B-Instruct-OpenOrca is a 1B parameter, causal decoder-only model based on [Falcon-RW-1B](https://huggingface.co/tiiuae/falcon-rw-1b) and finetuned on the [Open-Orca/SlimOrca](https://huggingface.co/datasets/Open-Orca/SlimOrca) dataset.
**✨Check out our new conversational model [Falcon-RW-1B-Chat](https://huggingface.co/ericzzz/falcon-rw-1b-chat)!✨**
**📊 Evaluation Results**
Falcon-RW-1B-Instruct-OpenOrca was the #1 ranking model (unfortunately not anymore) on [Open LLM Leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard) in ~1.5B parameters category! A detailed result can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_ericzzz__falcon-rw-1b-instruct-openorca).
| Metric | falcon-rw-1b-instruct-openorca | falcon-rw-1b |
|------------|-------------------------------:|-------------:|
| ARC | 34.56 | 35.07 |
| HellaSwag | 60.93 | 63.56 |
| MMLU | 28.77 | 25.28 |
| TruthfulQA | 37.42 | 35.96 |
| Winogrande | 60.69 | 62.04 |
| GSM8K | 3.41 | 0.53 |
| **Average**| **37.63** | **37.07** |
**🚀 Motivations**
1. To create a smaller, open-source, instruction-finetuned, ready-to-use model accessible for users with limited computational resources (lower-end consumer GPUs).
2. To harness the strength of Falcon-RW-1B, a competitive model in its own right, and enhance its capabilities with instruction finetuning.
## 📖 How to Use
The model operates with a structured prompt format, incorporating `<SYS>`, `<INST>`, and `<RESP>` tags to demarcate different parts of the input. The system message and instruction are placed within these tags, with the `<RESP>` tag triggering the model's response.
**📝 Example Code**
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
import transformers
import torch
model = 'ericzzz/falcon-rw-1b-instruct-openorca'
tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
'text-generation',
model=model,
tokenizer=tokenizer,
torch_dtype=torch.bfloat16,
device_map='auto',
)
system_message = 'You are a helpful assistant. Give short answers.'
instruction = 'What is AI? Give some examples.'
prompt = f'<SYS> {system_message} <INST> {instruction} <RESP> '
response = pipeline(
prompt,
max_length=200,
repetition_penalty=1.05
)
print(response[0]['generated_text'])
# AI, or Artificial Intelligence, refers to the ability of machines and software to perform tasks that require human intelligence, such as learning, reasoning, and problem-solving. It can be used in various fields like computer science, engineering, medicine, and more. Some common applications include image recognition, speech translation, and natural language processing.
```
## ⚠️ Limitations
This model may generate inaccurate or misleading information and is prone to hallucination, creating plausible but false narratives. It lacks the ability to discern factual content from fiction and may inadvertently produce biased, harmful or offensive content. Its understanding of complex, nuanced queries is limited. Users should be aware of this and verify any information obtained from the model.
The model is provided 'as is' without any warranties, and the creators are not liable for any damages arising from its use. Users are responsible for their interactions with the model.
## 📬 Contact
For further inquiries or feedback, please contact at [email protected].
## [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_ericzzz__falcon-rw-1b-instruct-openorca)
| Metric |Value|
|---------------------------------|----:|
|Avg. |37.63|
|AI2 Reasoning Challenge (25-Shot)|34.56|
|HellaSwag (10-Shot) |60.93|
|MMLU (5-Shot) |28.77|
|TruthfulQA (0-shot) |37.42|
|Winogrande (5-shot) |60.69|
|GSM8k (5-shot) | 3.41|
| [
"TRANSLATION"
] | Non_BioNLP |
fine-tuned/FiQA2018-256-24-gpt-4o-2024-05-13-256742 | fine-tuned | feature-extraction | [
"sentence-transformers",
"safetensors",
"bert",
"feature-extraction",
"sentence-similarity",
"mteb",
"en",
"dataset:fine-tuned/FiQA2018-256-24-gpt-4o-2024-05-13-256742",
"dataset:allenai/c4",
"license:apache-2.0",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,716,459,970,000 | 2024-05-23T10:26:22 | 9 | 0 | ---
datasets:
- fine-tuned/FiQA2018-256-24-gpt-4o-2024-05-13-256742
- allenai/c4
language:
- en
license: apache-2.0
pipeline_tag: feature-extraction
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- mteb
---
This model is a fine-tuned version of [**BAAI/bge-base-en-v1.5**](https://huggingface.co/BAAI/bge-base-en-v1.5) designed for the following use case:
custom
## How to Use
This model can be easily integrated into your NLP pipeline for tasks such as text classification, sentiment analysis, entity recognition, and more. Here's a simple example to get you started:
```python
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim
model = SentenceTransformer(
'fine-tuned/FiQA2018-256-24-gpt-4o-2024-05-13-256742',
trust_remote_code=True
)
embeddings = model.encode([
'first text to embed',
'second text to embed'
])
print(cos_sim(embeddings[0], embeddings[1]))
```
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
PragmaticPete/tinyqwen | PragmaticPete | text-generation | [
"transformers",
"safetensors",
"qwen2",
"text-generation",
"pretrained",
"conversational",
"en",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,718,651,742,000 | 2024-06-17T19:19:41 | 14 | 0 | ---
language:
- en
license: apache-2.0
pipeline_tag: text-generation
tags:
- pretrained
---
# Qwen2-0.5B
## Introduction
Qwen2 is the new series of Qwen large language models. For Qwen2, we release a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters, including a Mixture-of-Experts model. This repo contains the 0.5B Qwen2 base language model.
Compared with the state-of-the-art opensource language models, including the previous released Qwen1.5, Qwen2 has generally surpassed most opensource models and demonstrated competitiveness against proprietary models across a series of benchmarks targeting for language understanding, language generation, multilingual capability, coding, mathematics, reasoning, etc.
For more details, please refer to our [blog](https://qwenlm.github.io/blog/qwen2/), [GitHub](https://github.com/QwenLM/Qwen2), and [Documentation](https://qwen.readthedocs.io/en/latest/).
<br>
## Model Details
Qwen2 is a language model series including decoder language models of different model sizes. For each size, we release the base language model and the aligned chat model. It is based on the Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, etc. Additionally, we have an improved tokenizer adaptive to multiple natural languages and codes.
## Requirements
The code of Qwen2 has been in the latest Hugging face transformers and we advise you to install `transformers>=4.37.0`, or you might encounter the following error:
```
KeyError: 'qwen2'
```
## Usage
We do not advise you to use base language models for text generation. Instead, you can apply post-training, e.g., SFT, RLHF, continued pretraining, etc., on this model.
## Performance
The evaluation of base models mainly focuses on the model performance of natural language understanding, general question answering, coding, mathematics, scientific knowledge, reasoning, multilingual capability, etc.
The datasets for evaluation include:
**English Tasks**: MMLU (5-shot), MMLU-Pro (5-shot), GPQA (5shot), Theorem QA (5-shot), BBH (3-shot), HellaSwag (10-shot), Winogrande (5-shot), TruthfulQA (0-shot), ARC-C (25-shot)
**Coding Tasks**: EvalPlus (0-shot) (HumanEval, MBPP, HumanEval+, MBPP+), MultiPL-E (0-shot) (Python, C++, JAVA, PHP, TypeScript, C#, Bash, JavaScript)
**Math Tasks**: GSM8K (4-shot), MATH (4-shot)
**Chinese Tasks**: C-Eval(5-shot), CMMLU (5-shot)
**Multilingual Tasks**: Multi-Exam (M3Exam 5-shot, IndoMMLU 3-shot, ruMMLU 5-shot, mMMLU 5-shot), Multi-Understanding (BELEBELE 5-shot, XCOPA 5-shot, XWinograd 5-shot, XStoryCloze 0-shot, PAWS-X 5-shot), Multi-Mathematics (MGSM 8-shot), Multi-Translation (Flores-101 5-shot)
#### Qwen2-0.5B & Qwen2-1.5B performances
| Datasets | Phi-2 | Gemma-2B | MiniCPM | Qwen1.5-1.8B | Qwen2-0.5B | Qwen2-1.5B |
| :--------| :---------: | :------------: | :------------: |:------------: | :------------: | :------------: |
|#Non-Emb Params | 2.5B | 2.0B | 2.4B | 1.3B | 0.35B | 1.3B |
|MMLU | 52.7 | 42.3 | 53.5 | 46.8 | 45.4 | **56.5** |
|MMLU-Pro | - | 15.9 | - | - | 14.7 | 21.8 |
|Theorem QA | - | - | - |- | 8.9 | **15.0** |
|HumanEval | 47.6 | 22.0 |**50.0**| 20.1 | 22.0 | 31.1 |
|MBPP | **55.0** | 29.2 | 47.3 | 18.0 | 22.0 | 37.4 |
|GSM8K | 57.2 | 17.7 | 53.8 | 38.4 | 36.5 | **58.5** |
|MATH | 3.5 | 11.8 | 10.2 | 10.1 | 10.7 | **21.7** |
|BBH | **43.4** | 35.2 | 36.9 | 24.2 | 28.4 | 37.2 |
|HellaSwag | **73.1** | 71.4 | 68.3 | 61.4 | 49.3 | 66.6 |
|Winogrande | **74.4** | 66.8 | -| 60.3 | 56.8 | 66.2 |
|ARC-C | **61.1** | 48.5 | -| 37.9 | 31.5 | 43.9 |
|TruthfulQA | 44.5 | 33.1 | -| 39.4 | 39.7 | **45.9** |
|C-Eval | 23.4 | 28.0 | 51.1| 59.7 | 58.2 | **70.6** |
|CMMLU | 24.2 | - | 51.1 | 57.8 | 55.1 | **70.3** |
## Citation
If you find our work helpful, feel free to give us a cite.
```
@article{qwen2,
title={Qwen2 Technical Report},
year={2024}
}
```
| [
"QUESTION_ANSWERING",
"TRANSLATION"
] | Non_BioNLP |
Pclanglais/Larth-Mistral | Pclanglais | text-generation | [
"transformers",
"pytorch",
"mistral",
"text-generation",
"fr",
"license:cc-by-4.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"8-bit",
"bitsandbytes",
"region:us"
] | 1,696,941,413,000 | 2023-10-21T21:16:07 | 20 | 5 | ---
language:
- fr
library_name: transformers
license: cc-by-4.0
pipeline_tag: text-generation
widget:
- text: 'Answer in Etruscan: Who is the father of Lars?'
example_title: Lars
inference:
parameters:
temperature: 0.7
repetition_penalty: 1.2
---
Larth-Mistral is the first LLM based on the Etruscan language, fine-tuned on 1087 original inscriptions.
Larth-Mistral supports cross-linguistic instructions (question in English, answer in Etruscan) and automated translations. The formula to use are:
* *Answer in Etruscan: [Instruction in English]*
* *Translate in Etruscan: [Text in English]*
* *Translate in English: [Text in Etruscan]*
Larth-Mistral is not really an LLM *in* Etruscan but rather a poetic approximation of what LLM in Etruscan could be. Currently there is not even enough material to decipher the language, even less so for text generation.
The inscriptions used by Larth-Mistral were collected by the [Larth-Etruscan-NLP](https://github.com/GianlucaVico/Larth-Etruscan-NLP/tree/main) project
## Examples of instructions
**Who is the father of Lars?**
θre larsusas clareza ler θvele lupi lsveneii lcaii vnii lcaii avl ppa clesu claσ' menu hraχ
**What is the best recipe for an apple pie?**
θuva {m} : apuσ kaklaχ anaχ turuχ leθthturic lam θas acraχ clesinice leslumna τiχie huθ
**What is the best city of the dodecapolis?**
εις σθαχθρ τημενθ σπολισ δωδεκα πολισ σθαχθρ συφθ**
**Unexpectedly Larth-Mistral has switched to ancient greek.
| [
"TRANSLATION"
] | Non_BioNLP |
fine-tuned/SciFact-512-192-gpt-4o-2024-05-13-28032241 | fine-tuned | feature-extraction | [
"sentence-transformers",
"safetensors",
"bert",
"feature-extraction",
"sentence-similarity",
"mteb",
"en",
"dataset:fine-tuned/SciFact-512-192-gpt-4o-2024-05-13-28032241",
"dataset:allenai/c4",
"license:apache-2.0",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,716,922,458,000 | 2024-05-28T18:54:49 | 6 | 0 | ---
datasets:
- fine-tuned/SciFact-512-192-gpt-4o-2024-05-13-28032241
- allenai/c4
language:
- en
- en
license: apache-2.0
pipeline_tag: feature-extraction
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- mteb
---
This model is a fine-tuned version of [**BAAI/bge-large-en-v1.5**](https://huggingface.co/BAAI/bge-large-en-v1.5) designed for the following use case:
None
## How to Use
This model can be easily integrated into your NLP pipeline for tasks such as text classification, sentiment analysis, entity recognition, and more. Here's a simple example to get you started:
```python
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import cos_sim
model = SentenceTransformer(
'fine-tuned/SciFact-512-192-gpt-4o-2024-05-13-28032241',
trust_remote_code=True
)
embeddings = model.encode([
'first text to embed',
'second text to embed'
])
print(cos_sim(embeddings[0], embeddings[1]))
```
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
pEpOo/catastrophy8 | pEpOo | text-classification | [
"setfit",
"safetensors",
"mpnet",
"sentence-transformers",
"text-classification",
"generated_from_setfit_trainer",
"arxiv:2209.11055",
"base_model:sentence-transformers/all-mpnet-base-v2",
"base_model:finetune:sentence-transformers/all-mpnet-base-v2",
"model-index",
"region:us"
] | 1,702,908,844,000 | 2023-12-18T14:14:25 | 50 | 0 | ---
base_model: sentence-transformers/all-mpnet-base-v2
library_name: setfit
metrics:
- accuracy
pipeline_tag: text-classification
tags:
- setfit
- sentence-transformers
- text-classification
- generated_from_setfit_trainer
widget:
- text: "Rly tragedy in MP: Some live to recount horror: \x89ÛÏWhen I saw coaches\
\ of my train plunging into water I called my daughters and said t..."
- text: You must be annihilated!
- text: 'Severe Thunderstorms and Flash Flooding Possible in the Mid-South and Midwest
http://t.co/uAhIcWpIh4 #WEATHER #ENVIRONMENT #CLIMATE #NATURE'
- text: 'everyone''s wonder who will win and I''m over here wondering are those grapes
real ?????? #BB17'
- text: i swea it feels like im about to explode ??
inference: true
model-index:
- name: SetFit with sentence-transformers/all-mpnet-base-v2
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: Unknown
type: unknown
split: test
metrics:
- type: accuracy
value: 0.9203152364273205
name: Accuracy
---
# SetFit with sentence-transformers/all-mpnet-base-v2
This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [sentence-transformers/all-mpnet-base-v2](https://huggingface.co/sentence-transformers/all-mpnet-base-v2) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification.
The model has been trained using an efficient few-shot learning technique that involves:
1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.
## Model Details
### Model Description
- **Model Type:** SetFit
- **Sentence Transformer body:** [sentence-transformers/all-mpnet-base-v2](https://huggingface.co/sentence-transformers/all-mpnet-base-v2)
- **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
- **Maximum Sequence Length:** 384 tokens
- **Number of Classes:** 2 classes
<!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)
- **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
### Model Labels
| Label | Examples |
|:------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| 0 | <ul><li>'To fight bioterrorism sir.'</li><li>'85V-265V 10W LED Warm White Light Motion Sensor Outdoor Flood Light PIR Lamp AUC http://t.co/NJVPXzMj5V http://t.co/Ijd7WzV5t9'</li><li>'Photo: referencereference: xekstrin: I THOUGHT THE NOSTRILS WERE EYES AND I ALMOST CRIED FROM FEAR partake... http://t.co/O7yYjLuKfJ'</li></ul> |
| 1 | <ul><li>'Police officer wounded suspect dead after exchanging shots: RICHMOND Va. (AP) \x89ÛÓ A Richmond police officer wa... http://t.co/Y0qQS2L7bS'</li><li>"There's a weird siren going off here...I hope Hunterston isn't in the process of blowing itself to smithereens..."</li><li>'Iranian warship points weapon at American helicopter... http://t.co/cgFZk8Ha1R'</li></ul> |
## Evaluation
### Metrics
| Label | Accuracy |
|:--------|:---------|
| **all** | 0.9203 |
## Uses
### Direct Use for Inference
First install the SetFit library:
```bash
pip install setfit
```
Then you can load this model and run inference.
```python
from setfit import SetFitModel
# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("pEpOo/catastrophy8")
# Run inference
preds = model("You must be annihilated!")
```
<!--
### Downstream Use
*List how someone could finetune this model on their own dataset.*
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Set Metrics
| Training set | Min | Median | Max |
|:-------------|:----|:--------|:----|
| Word count | 1 | 14.5506 | 54 |
| Label | Training Sample Count |
|:------|:----------------------|
| 0 | 438 |
| 1 | 323 |
### Training Hyperparameters
- batch_size: (20, 20)
- num_epochs: (1, 1)
- max_steps: -1
- sampling_strategy: oversampling
- num_iterations: 20
- body_learning_rate: (2e-05, 2e-05)
- head_learning_rate: 2e-05
- loss: CosineSimilarityLoss
- distance_metric: cosine_distance
- margin: 0.25
- end_to_end: False
- use_amp: False
- warmup_proportion: 0.1
- seed: 42
- eval_max_steps: -1
- load_best_model_at_end: False
### Training Results
| Epoch | Step | Training Loss | Validation Loss |
|:------:|:-----:|:-------------:|:---------------:|
| 0.0001 | 1 | 0.3847 | - |
| 0.0044 | 50 | 0.3738 | - |
| 0.0088 | 100 | 0.2274 | - |
| 0.0131 | 150 | 0.2747 | - |
| 0.0175 | 200 | 0.2251 | - |
| 0.0219 | 250 | 0.2562 | - |
| 0.0263 | 300 | 0.2623 | - |
| 0.0307 | 350 | 0.1904 | - |
| 0.0350 | 400 | 0.2314 | - |
| 0.0394 | 450 | 0.1669 | - |
| 0.0438 | 500 | 0.1135 | - |
| 0.0482 | 550 | 0.1489 | - |
| 0.0525 | 600 | 0.1907 | - |
| 0.0569 | 650 | 0.1728 | - |
| 0.0613 | 700 | 0.125 | - |
| 0.0657 | 750 | 0.109 | - |
| 0.0701 | 800 | 0.0968 | - |
| 0.0744 | 850 | 0.2101 | - |
| 0.0788 | 900 | 0.1974 | - |
| 0.0832 | 950 | 0.1986 | - |
| 0.0876 | 1000 | 0.0747 | - |
| 0.0920 | 1050 | 0.1117 | - |
| 0.0963 | 1100 | 0.1092 | - |
| 0.1007 | 1150 | 0.1582 | - |
| 0.1051 | 1200 | 0.1243 | - |
| 0.1095 | 1250 | 0.2873 | - |
| 0.1139 | 1300 | 0.2415 | - |
| 0.1182 | 1350 | 0.1264 | - |
| 0.1226 | 1400 | 0.127 | - |
| 0.1270 | 1450 | 0.1308 | - |
| 0.1314 | 1500 | 0.0669 | - |
| 0.1358 | 1550 | 0.1218 | - |
| 0.1401 | 1600 | 0.114 | - |
| 0.1445 | 1650 | 0.0612 | - |
| 0.1489 | 1700 | 0.0527 | - |
| 0.1533 | 1750 | 0.1421 | - |
| 0.1576 | 1800 | 0.0048 | - |
| 0.1620 | 1850 | 0.0141 | - |
| 0.1664 | 1900 | 0.0557 | - |
| 0.1708 | 1950 | 0.0206 | - |
| 0.1752 | 2000 | 0.1171 | - |
| 0.1795 | 2050 | 0.0968 | - |
| 0.1839 | 2100 | 0.0243 | - |
| 0.1883 | 2150 | 0.0233 | - |
| 0.1927 | 2200 | 0.0738 | - |
| 0.1971 | 2250 | 0.0071 | - |
| 0.2014 | 2300 | 0.0353 | - |
| 0.2058 | 2350 | 0.0602 | - |
| 0.2102 | 2400 | 0.003 | - |
| 0.2146 | 2450 | 0.0625 | - |
| 0.2190 | 2500 | 0.0173 | - |
| 0.2233 | 2550 | 0.1017 | - |
| 0.2277 | 2600 | 0.0582 | - |
| 0.2321 | 2650 | 0.0437 | - |
| 0.2365 | 2700 | 0.104 | - |
| 0.2408 | 2750 | 0.0156 | - |
| 0.2452 | 2800 | 0.0034 | - |
| 0.2496 | 2850 | 0.0343 | - |
| 0.2540 | 2900 | 0.1106 | - |
| 0.2584 | 2950 | 0.001 | - |
| 0.2627 | 3000 | 0.004 | - |
| 0.2671 | 3050 | 0.0074 | - |
| 0.2715 | 3100 | 0.0849 | - |
| 0.2759 | 3150 | 0.0009 | - |
| 0.2803 | 3200 | 0.0379 | - |
| 0.2846 | 3250 | 0.0109 | - |
| 0.2890 | 3300 | 0.0019 | - |
| 0.2934 | 3350 | 0.0154 | - |
| 0.2978 | 3400 | 0.0017 | - |
| 0.3022 | 3450 | 0.0003 | - |
| 0.3065 | 3500 | 0.0002 | - |
| 0.3109 | 3550 | 0.0025 | - |
| 0.3153 | 3600 | 0.0123 | - |
| 0.3197 | 3650 | 0.0007 | - |
| 0.3240 | 3700 | 0.0534 | - |
| 0.3284 | 3750 | 0.0004 | - |
| 0.3328 | 3800 | 0.0084 | - |
| 0.3372 | 3850 | 0.0088 | - |
| 0.3416 | 3900 | 0.0201 | - |
| 0.3459 | 3950 | 0.0002 | - |
| 0.3503 | 4000 | 0.0102 | - |
| 0.3547 | 4050 | 0.0043 | - |
| 0.3591 | 4100 | 0.0124 | - |
| 0.3635 | 4150 | 0.0845 | - |
| 0.3678 | 4200 | 0.0002 | - |
| 0.3722 | 4250 | 0.0014 | - |
| 0.3766 | 4300 | 0.1131 | - |
| 0.3810 | 4350 | 0.0612 | - |
| 0.3854 | 4400 | 0.0577 | - |
| 0.3897 | 4450 | 0.0235 | - |
| 0.3941 | 4500 | 0.0156 | - |
| 0.3985 | 4550 | 0.0078 | - |
| 0.4029 | 4600 | 0.0356 | - |
| 0.4073 | 4650 | 0.0595 | - |
| 0.4116 | 4700 | 0.0001 | - |
| 0.4160 | 4750 | 0.0018 | - |
| 0.4204 | 4800 | 0.0013 | - |
| 0.4248 | 4850 | 0.0008 | - |
| 0.4291 | 4900 | 0.0832 | - |
| 0.4335 | 4950 | 0.0083 | - |
| 0.4379 | 5000 | 0.0007 | - |
| 0.4423 | 5050 | 0.0417 | - |
| 0.4467 | 5100 | 0.0001 | - |
| 0.4510 | 5150 | 0.0218 | - |
| 0.4554 | 5200 | 0.0001 | - |
| 0.4598 | 5250 | 0.0012 | - |
| 0.4642 | 5300 | 0.0002 | - |
| 0.4686 | 5350 | 0.0006 | - |
| 0.4729 | 5400 | 0.0223 | - |
| 0.4773 | 5450 | 0.0612 | - |
| 0.4817 | 5500 | 0.0004 | - |
| 0.4861 | 5550 | 0.0 | - |
| 0.4905 | 5600 | 0.0007 | - |
| 0.4948 | 5650 | 0.0007 | - |
| 0.4992 | 5700 | 0.0116 | - |
| 0.5036 | 5750 | 0.0262 | - |
| 0.5080 | 5800 | 0.0336 | - |
| 0.5123 | 5850 | 0.026 | - |
| 0.5167 | 5900 | 0.0004 | - |
| 0.5211 | 5950 | 0.0001 | - |
| 0.5255 | 6000 | 0.0001 | - |
| 0.5299 | 6050 | 0.0001 | - |
| 0.5342 | 6100 | 0.0029 | - |
| 0.5386 | 6150 | 0.0001 | - |
| 0.5430 | 6200 | 0.0699 | - |
| 0.5474 | 6250 | 0.0262 | - |
| 0.5518 | 6300 | 0.0269 | - |
| 0.5561 | 6350 | 0.0002 | - |
| 0.5605 | 6400 | 0.0666 | - |
| 0.5649 | 6450 | 0.0209 | - |
| 0.5693 | 6500 | 0.0003 | - |
| 0.5737 | 6550 | 0.0001 | - |
| 0.5780 | 6600 | 0.0115 | - |
| 0.5824 | 6650 | 0.0003 | - |
| 0.5868 | 6700 | 0.0001 | - |
| 0.5912 | 6750 | 0.0056 | - |
| 0.5956 | 6800 | 0.0603 | - |
| 0.5999 | 6850 | 0.0002 | - |
| 0.6043 | 6900 | 0.0003 | - |
| 0.6087 | 6950 | 0.0092 | - |
| 0.6131 | 7000 | 0.0562 | - |
| 0.6174 | 7050 | 0.0408 | - |
| 0.6218 | 7100 | 0.0001 | - |
| 0.6262 | 7150 | 0.0035 | - |
| 0.6306 | 7200 | 0.0337 | - |
| 0.6350 | 7250 | 0.0024 | - |
| 0.6393 | 7300 | 0.0005 | - |
| 0.6437 | 7350 | 0.0001 | - |
| 0.6481 | 7400 | 0.0 | - |
| 0.6525 | 7450 | 0.0001 | - |
| 0.6569 | 7500 | 0.0002 | - |
| 0.6612 | 7550 | 0.0004 | - |
| 0.6656 | 7600 | 0.0125 | - |
| 0.6700 | 7650 | 0.0005 | - |
| 0.6744 | 7700 | 0.0157 | - |
| 0.6788 | 7750 | 0.0055 | - |
| 0.6831 | 7800 | 0.0 | - |
| 0.6875 | 7850 | 0.0053 | - |
| 0.6919 | 7900 | 0.0 | - |
| 0.6963 | 7950 | 0.0002 | - |
| 0.7006 | 8000 | 0.0002 | - |
| 0.7050 | 8050 | 0.0001 | - |
| 0.7094 | 8100 | 0.0001 | - |
| 0.7138 | 8150 | 0.0001 | - |
| 0.7182 | 8200 | 0.0007 | - |
| 0.7225 | 8250 | 0.0002 | - |
| 0.7269 | 8300 | 0.0001 | - |
| 0.7313 | 8350 | 0.0 | - |
| 0.7357 | 8400 | 0.0156 | - |
| 0.7401 | 8450 | 0.0098 | - |
| 0.7444 | 8500 | 0.0 | - |
| 0.7488 | 8550 | 0.0001 | - |
| 0.7532 | 8600 | 0.0042 | - |
| 0.7576 | 8650 | 0.0 | - |
| 0.7620 | 8700 | 0.0 | - |
| 0.7663 | 8750 | 0.0056 | - |
| 0.7707 | 8800 | 0.0 | - |
| 0.7751 | 8850 | 0.0 | - |
| 0.7795 | 8900 | 0.013 | - |
| 0.7839 | 8950 | 0.0 | - |
| 0.7882 | 9000 | 0.0001 | - |
| 0.7926 | 9050 | 0.0 | - |
| 0.7970 | 9100 | 0.0 | - |
| 0.8014 | 9150 | 0.0 | - |
| 0.8057 | 9200 | 0.0 | - |
| 0.8101 | 9250 | 0.0 | - |
| 0.8145 | 9300 | 0.0007 | - |
| 0.8189 | 9350 | 0.0 | - |
| 0.8233 | 9400 | 0.0002 | - |
| 0.8276 | 9450 | 0.0 | - |
| 0.8320 | 9500 | 0.0 | - |
| 0.8364 | 9550 | 0.0089 | - |
| 0.8408 | 9600 | 0.0001 | - |
| 0.8452 | 9650 | 0.0 | - |
| 0.8495 | 9700 | 0.0 | - |
| 0.8539 | 9750 | 0.0 | - |
| 0.8583 | 9800 | 0.0565 | - |
| 0.8627 | 9850 | 0.0161 | - |
| 0.8671 | 9900 | 0.0 | - |
| 0.8714 | 9950 | 0.0246 | - |
| 0.8758 | 10000 | 0.0 | - |
| 0.8802 | 10050 | 0.0 | - |
| 0.8846 | 10100 | 0.012 | - |
| 0.8889 | 10150 | 0.0 | - |
| 0.8933 | 10200 | 0.0 | - |
| 0.8977 | 10250 | 0.0 | - |
| 0.9021 | 10300 | 0.0 | - |
| 0.9065 | 10350 | 0.0 | - |
| 0.9108 | 10400 | 0.0 | - |
| 0.9152 | 10450 | 0.0 | - |
| 0.9196 | 10500 | 0.0 | - |
| 0.9240 | 10550 | 0.0023 | - |
| 0.9284 | 10600 | 0.0 | - |
| 0.9327 | 10650 | 0.0006 | - |
| 0.9371 | 10700 | 0.0 | - |
| 0.9415 | 10750 | 0.0 | - |
| 0.9459 | 10800 | 0.0 | - |
| 0.9503 | 10850 | 0.0 | - |
| 0.9546 | 10900 | 0.0 | - |
| 0.9590 | 10950 | 0.0243 | - |
| 0.9634 | 11000 | 0.0107 | - |
| 0.9678 | 11050 | 0.0001 | - |
| 0.9721 | 11100 | 0.0 | - |
| 0.9765 | 11150 | 0.0 | - |
| 0.9809 | 11200 | 0.0274 | - |
| 0.9853 | 11250 | 0.0 | - |
| 0.9897 | 11300 | 0.0 | - |
| 0.9940 | 11350 | 0.0 | - |
| 0.9984 | 11400 | 0.0 | - |
| 0.0007 | 1 | 0.2021 | - |
| 0.0329 | 50 | 0.1003 | - |
| 0.0657 | 100 | 0.2282 | - |
| 0.0986 | 150 | 0.0507 | - |
| 0.1314 | 200 | 0.046 | - |
| 0.1643 | 250 | 0.0001 | - |
| 0.1971 | 300 | 0.0495 | - |
| 0.2300 | 350 | 0.0031 | - |
| 0.2628 | 400 | 0.0004 | - |
| 0.2957 | 450 | 0.0002 | - |
| 0.3285 | 500 | 0.0 | - |
| 0.3614 | 550 | 0.0 | - |
| 0.3942 | 600 | 0.0 | - |
| 0.4271 | 650 | 0.0001 | - |
| 0.4599 | 700 | 0.0 | - |
| 0.4928 | 750 | 0.0 | - |
| 0.5256 | 800 | 0.0 | - |
| 0.5585 | 850 | 0.0 | - |
| 0.5913 | 900 | 0.0001 | - |
| 0.6242 | 950 | 0.0 | - |
| 0.6570 | 1000 | 0.0001 | - |
| 0.6899 | 1050 | 0.0 | - |
| 0.7227 | 1100 | 0.0 | - |
| 0.7556 | 1150 | 0.0 | - |
| 0.7884 | 1200 | 0.0 | - |
| 0.8213 | 1250 | 0.0 | - |
| 0.8541 | 1300 | 0.0 | - |
| 0.8870 | 1350 | 0.0 | - |
| 0.9198 | 1400 | 0.0 | - |
| 0.9527 | 1450 | 0.0001 | - |
| 0.9855 | 1500 | 0.0 | - |
### Framework Versions
- Python: 3.10.12
- SetFit: 1.0.1
- Sentence Transformers: 2.2.2
- Transformers: 4.35.2
- PyTorch: 2.1.0+cu121
- Datasets: 2.15.0
- Tokenizers: 0.15.0
## Citation
### BibTeX
```bibtex
@article{https://doi.org/10.48550/arxiv.2209.11055,
doi = {10.48550/ARXIV.2209.11055},
url = {https://arxiv.org/abs/2209.11055},
author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Efficient Few-Shot Learning Without Prompts},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
Anjaan-Khadka/Nepali-Summarization | Anjaan-Khadka | summarization | [
"transformers",
"pytorch",
"mt5",
"text2text-generation",
"summarization",
"mT5",
"ne",
"dataset:csebuetnlp/xlsum",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,677,152,698,000 | 2023-03-17T08:45:04 | 21 | 0 | ---
datasets:
- csebuetnlp/xlsum
language:
- ne
tags:
- summarization
- mT5
widget:
- text: तीन नगरपालिकालाई समेटेर भेरी किनारमा बन्न थालेको आधुनिक नमुना सहरको काम तीव्र
गतिमा अघि बढेको छ । भेरीगंगा, गुर्भाकोट र लेकबेंसी नगरपालिकामा बन्न थालेको भेरीगंगा
उपत्यका नमुना आधुनिक सहर निर्माण हुन लागेको हो । यसले नदी वारि र पारिको ४ सय ६०
वर्ग किलोमिटर क्षेत्रलाई समेट्नेछ ।
model-index:
- name: Anjaan-Khadka/summarization_nepali
results:
- task:
type: summarization
name: Summarization
dataset:
name: xsum
type: xsum
config: default
split: test
metrics:
- type: rouge
value: 36.5002
name: ROUGE-1
verified: false
---
# adaptation of mT5-multilingual-XLSum for Nepali Lnaguage
This repository contains adapted version of mT5-multilinguag-XLSum for Single Language (Nepali). View original [mT5-multilinguag-XLSum model](https://huggingface.co/csebuetnlp/mT5_multilingual_XLSum)
## Using this model in `transformers` (tested on 4.11.0.dev0)
```python
import re
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
article_text = " तीन नगरपालिकालाई समेटेर भेरी किनारमा बन्न थालेको आधुनिक नमुना सहरको काम तीव्र गतिमा अघि बढेको छ । भेरीगंगा, गुर्भाकोट र लेकबेंसी नगरपालिकामा बन्न थालेको भेरीगंगा उपत्यका नमुना आधुनिक सहर निर्माण हुन लागेको हो । यसले नदी वारि र पारिको ४ सय ६० वर्ग किलोमिटर क्षेत्रलाई समेट्नेछ ।"
model_name = "Anjaan-Khadka/summarization_nepali"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
input_ids = tokenizer(
(article_text),
return_tensors="pt",
padding="max_length",
truncation=True,
max_length=512
)["input_ids"]
output_ids = model.generate(
input_ids=input_ids,
max_length=84,
no_repeat_ngram_size=2,
num_beams=4
)[0]
summary = tokenizer.decode(
output_ids,
skip_special_tokens=True,
clean_up_tokenization_spaces=False
)
print(summary)
``` | [
"SUMMARIZATION"
] | Non_BioNLP |
sndsabin/fake-news-classifier | sndsabin | null | [
"license:gpl-3.0",
"region:us"
] | 1,648,716,829,000 | 2022-04-07T08:58:17 | 0 | 0 | ---
license: gpl-3.0
---
**Fake News Classifier**: Text classification model to detect fake news articles!
**Dataset**: [Kaggle Fake and real news dataset](https://www.kaggle.com/datasets/clmentbisaillon/fake-and-real-news-dataset)
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF | TheBloke | text-generation | [
"transformers",
"gguf",
"solar",
"finetune",
"dpo",
"Instruct",
"augmentation",
"german",
"text-generation",
"en",
"de",
"dataset:argilla/distilabel-math-preference-dpo",
"base_model:fblgit/LUNA-SOLARkrautLM-Instruct",
"base_model:quantized:fblgit/LUNA-SOLARkrautLM-Instruct",
"license:cc-by-nc-4.0",
"region:us",
"conversational"
] | 1,703,336,543,000 | 2023-12-23T13:08:59 | 368 | 4 | ---
base_model: fblgit/LUNA-SOLARkrautLM-Instruct
datasets:
- argilla/distilabel-math-preference-dpo
language:
- en
- de
library_name: transformers
license: cc-by-nc-4.0
model_name: Luna SOLARkrautLM Instruct
pipeline_tag: text-generation
tags:
- finetune
- dpo
- Instruct
- augmentation
- german
inference: false
model_creator: FBL
model_type: solar
prompt_template: '<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
'
quantized_by: TheBloke
---
<!-- markdownlint-disable MD041 -->
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Luna SOLARkrautLM Instruct - GGUF
- Model creator: [FBL](https://huggingface.co/fblgit)
- Original model: [Luna SOLARkrautLM Instruct](https://huggingface.co/fblgit/LUNA-SOLARkrautLM-Instruct)
<!-- description start -->
## Description
This repo contains GGUF format model files for [FBL's Luna SOLARkrautLM Instruct](https://huggingface.co/fblgit/LUNA-SOLARkrautLM-Instruct).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
<!-- description end -->
<!-- README_GGUF.md-about-gguf start -->
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [GPT4All](https://gpt4all.io/index.html), a free and open source local running GUI, supporting Windows, Linux and macOS with full GPU accel.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration. Linux available, in beta as of 27/11/2023.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server. Note, as of time of writing (November 27th 2023), ctransformers has not been updated in a long time and does not support many recent models.
<!-- README_GGUF.md-about-gguf end -->
<!-- repositories-available start -->
## Repositories available
* [AWQ model(s) for GPU inference.](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-AWQ)
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF)
* [FBL's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/fblgit/LUNA-SOLARkrautLM-Instruct)
<!-- repositories-available end -->
<!-- prompt-template start -->
## Prompt template: ChatML
```
<|im_start|>system
{system_message}<|im_end|>
<|im_start|>user
{prompt}<|im_end|>
<|im_start|>assistant
```
<!-- prompt-template end -->
<!-- compatibility_gguf start -->
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
<details>
<summary>Click to see details</summary>
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
</details>
<!-- compatibility_gguf end -->
<!-- README_GGUF.md-provided-files start -->
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [luna-solarkrautlm-instruct.Q2_K.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q2_K.gguf) | Q2_K | 2 | 4.55 GB| 7.05 GB | smallest, significant quality loss - not recommended for most purposes |
| [luna-solarkrautlm-instruct.Q3_K_S.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q3_K_S.gguf) | Q3_K_S | 3 | 4.66 GB| 7.16 GB | very small, high quality loss |
| [luna-solarkrautlm-instruct.Q3_K_M.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q3_K_M.gguf) | Q3_K_M | 3 | 5.19 GB| 7.69 GB | very small, high quality loss |
| [luna-solarkrautlm-instruct.Q3_K_L.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q3_K_L.gguf) | Q3_K_L | 3 | 5.65 GB| 8.15 GB | small, substantial quality loss |
| [luna-solarkrautlm-instruct.Q4_0.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q4_0.gguf) | Q4_0 | 4 | 6.07 GB| 8.57 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [luna-solarkrautlm-instruct.Q4_K_S.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q4_K_S.gguf) | Q4_K_S | 4 | 6.10 GB| 8.60 GB | small, greater quality loss |
| [luna-solarkrautlm-instruct.Q4_K_M.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q4_K_M.gguf) | Q4_K_M | 4 | 6.46 GB| 8.96 GB | medium, balanced quality - recommended |
| [luna-solarkrautlm-instruct.Q5_0.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q5_0.gguf) | Q5_0 | 5 | 7.40 GB| 9.90 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [luna-solarkrautlm-instruct.Q5_K_S.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q5_K_S.gguf) | Q5_K_S | 5 | 7.40 GB| 9.90 GB | large, low quality loss - recommended |
| [luna-solarkrautlm-instruct.Q5_K_M.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q5_K_M.gguf) | Q5_K_M | 5 | 7.60 GB| 10.10 GB | large, very low quality loss - recommended |
| [luna-solarkrautlm-instruct.Q6_K.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q6_K.gguf) | Q6_K | 6 | 8.81 GB| 11.31 GB | very large, extremely low quality loss |
| [luna-solarkrautlm-instruct.Q8_0.gguf](https://huggingface.co/TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF/blob/main/luna-solarkrautlm-instruct.Q8_0.gguf) | Q8_0 | 8 | 11.40 GB| 13.90 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
<!-- README_GGUF.md-provided-files end -->
<!-- README_GGUF.md-how-to-download start -->
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF and below it, a specific filename to download, such as: luna-solarkrautlm-instruct.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF luna-solarkrautlm-instruct.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
<details>
<summary>More advanced huggingface-cli download usage (click to read)</summary>
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/LUNA-SOLARkrautLM-Instruct-GGUF luna-solarkrautlm-instruct.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
</details>
<!-- README_GGUF.md-how-to-download end -->
<!-- README_GGUF.md-how-to-run start -->
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 35 -m luna-solarkrautlm-instruct.Q4_K_M.gguf --color -c 4096 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 4096` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the `-p <PROMPT>` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
### How to load this model in Python code, using llama-cpp-python
For full documentation, please see: [llama-cpp-python docs](https://abetlen.github.io/llama-cpp-python/).
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python
```
#### Simple llama-cpp-python example code
```python
from llama_cpp import Llama
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = Llama(
model_path="./luna-solarkrautlm-instruct.Q4_K_M.gguf", # Download the model file first
n_ctx=4096, # The max sequence length to use - note that longer sequence lengths require much more resources
n_threads=8, # The number of CPU threads to use, tailor to your system and the resulting performance
n_gpu_layers=35 # The number of layers to offload to GPU, if you have GPU acceleration available
)
# Simple inference example
output = llm(
"<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant", # Prompt
max_tokens=512, # Generate up to 512 tokens
stop=["</s>"], # Example stop token - not necessarily correct for this specific model! Please check before using.
echo=True # Whether to echo the prompt
)
# Chat Completion API
llm = Llama(model_path="./luna-solarkrautlm-instruct.Q4_K_M.gguf", chat_format="llama-2") # Set chat_format according to the model you are using
llm.create_chat_completion(
messages = [
{"role": "system", "content": "You are a story writing assistant."},
{
"role": "user",
"content": "Write a story about llamas."
}
]
)
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
<!-- README_GGUF.md-how-to-run end -->
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Michael Levine, 阿明, Trailburnt, Nikolai Manek, John Detwiler, Randy H, Will Dee, Sebastain Graf, NimbleBox.ai, Eugene Pentland, Emad Mostaque, Ai Maven, Jim Angel, Jeff Scroggin, Michael Davis, Manuel Alberto Morcote, Stephen Murray, Robert, Justin Joy, Luke @flexchar, Brandon Frisco, Elijah Stavena, S_X, Dan Guido, Undi ., Komninos Chatzipapas, Shadi, theTransient, Lone Striker, Raven Klaugh, jjj, Cap'n Zoog, Michel-Marie MAUDET (LINAGORA), Matthew Berman, David, Fen Risland, Omer Bin Jawed, Luke Pendergrass, Kalila, OG, Erik Bjäreholt, Rooh Singh, Joseph William Delisle, Dan Lewis, TL, John Villwock, AzureBlack, Brad, Pedro Madruga, Caitlyn Gatomon, K, jinyuan sun, Mano Prime, Alex, Jeffrey Morgan, Alicia Loh, Illia Dulskyi, Chadd, transmissions 11, fincy, Rainer Wilmers, ReadyPlayerEmma, knownsqashed, Mandus, biorpg, Deo Leter, Brandon Phillips, SuperWojo, Sean Connelly, Iucharbius, Jack West, Harry Royden McLaughlin, Nicholas, terasurfer, Vitor Caleffi, Duane Dunston, Johann-Peter Hartmann, David Ziegler, Olakabola, Ken Nordquist, Trenton Dambrowitz, Tom X Nguyen, Vadim, Ajan Kanaga, Leonard Tan, Clay Pascal, Alexandros Triantafyllidis, JM33133, Xule, vamX, ya boyyy, subjectnull, Talal Aujan, Alps Aficionado, wassieverse, Ari Malik, James Bentley, Woland, Spencer Kim, Michael Dempsey, Fred von Graf, Elle, zynix, William Richards, Stanislav Ovsiannikov, Edmond Seymore, Jonathan Leane, Martin Kemka, usrbinkat, Enrico Ros
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
<!-- original-model-card start -->
# Original model card: FBL's Luna SOLARkrautLM Instruct
## VAGO solutions LUNA-SOLARkrautLM-Instruct
Introducing **LUNA-SOLARkrautLM-Instruct** – a UNA-Sauerkraut version of the powerful [upstage/SOLAR-10.7B-Instruct-v1.0](https://huggingface.co/upstage/SOLAR-10.7B-Instruct-v1.0) !
Aligned with **DPO** and tamed with **UNA**.
# Table of Contents
1. [Overview of all LUNA-SOLARkrautLM-Instruct models](#all-sauerkrautlm-solar-instruct-models)
2. [Model Details](#model-details)
- [Prompt template](#prompt-template)
- [Training Dataset](#training-dataset)
- [Data Contamination Test](#data-contamination-test-results)
3. [Evaluation](#evaluation)
5. [Disclaimer](#disclaimer)
6. [Contact](#contact)
7. [Collaborations](#collaborations)
8. [Acknowledgement](#acknowledgement)
## Model Details
**LUNA-SOLARkrautLM-Instruct**
- **Model Type:** LUNA-SOLARkrautLM-Instruct is a UNA Model based on [fblgit/UNA-SOLAR-10.7B-Instruct-v1.0](https://huggingface.co/fblgit/UNA-SOLAR-10.7B-Instruct-v1.0) and the powerful set of [SauerkrautLM-SOLAR-Instruct](https://huggingface.co/VAGOsolutions/SauerkrautLM-SOLAR-Instruct/)
- **Language(s):** English, German
- **License:** cc-by-nc-4.0
- **Contact:** [Website](https://vago-solutions.de/#Kontakt) [David Golchinfar](mailto:[email protected]) [Juanako.AI - UNA](mailto:[email protected])
### Training Dataset:
LUNA-SOLARkrautLM-Instruct was trained with mix of German data augmentation and translated data.
Aligned through **DPO** with our **new German SauerkrautLM-DPO dataset** based on parts of the SFT SauerkrautLM dataset
as chosen answers and [Sauerkraut-7b-HerO](https://huggingface.co/VAGOsolutions/SauerkrautLM-7b-HerO) as rejected answers. Added with additional **translated Parts of the [HuggingFaceH4/ultrafeedback_binarized](https://huggingface.co/datasets/HuggingFaceH4/ultrafeedback_binarized)** (Our dataset do not contain any TruthfulQA prompts - check Data Contamination Test Results) and **[argilla/distilabel-math-preference-dpo](https://huggingface.co/datasets/argilla/distilabel-math-preference-dpo).**
We found, that only a simple translation of training data can lead to unnatural German phrasings.
Data augmentation techniques were used to grant grammatical, syntactical correctness and a more natural German wording in our training data.
We improved the German language skills on this model. Nevertheless, certain formulations may occur that are not entirely correct.
### Data Contamination Test Results
Some models on the HuggingFace leaderboard had problems with wrong data getting mixed in.
We checked our SauerkrautLM-DPO dataset with a special test [1] on this model as target model and upstage/SOLAR-10.7B-Instruct-v1.0 as reference model.
The HuggingFace team used the same methods [2, 3].
Our results, with `result < 0.1, %:` being well below 0.9, indicate that our dataset is free from contamination.
*The data contamination test results of HellaSwag and Winograde will be added once [1] supports them.*
| Dataset | ARC | MMLU | TruthfulQA | GSM8K |
|------------------------------|-------|-------|-------|-------|
| **SauerkrautLM-DPO**| result < 0.1, %: 0.0 |result < 0.1, %: 0.09 | result < 0.1, %: 0.13 | result < 0.1, %: 0.16 |
[1] https://github.com/swj0419/detect-pretrain-code-contamination
[2] https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard/discussions/474#657f2245365456e362412a06
[3] https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard/discussions/265#657b6debf81f6b44b8966230
### Prompt Template:
```
<|im_start|>system
Du bist LUNA-SOLARkrautLM, ein großes Sprachmodell, das höflich und kompetent antwortet.<|im_end|>
<|im_start|>user
Wie geht es dir?<|im_end|>
<|im_start|>assistant
```
```
### User:
Hello, how are you?
### Assistant:
Hi there! I am an AI language model, so I don't have personal feelings or emotions in the traditional sense. However, I can assure you that my systems and processes are functioning well at this moment, allowing me to provide helpful responses for your queries.
How may I assist you today?
```
## Evaluation
```
hf (pretrained=fblgit/LUNA-SOLARkrautLM-Instruct), gen_kwargs: (), limit: None, num_fewshot: 5, batch_size: auto
|Tasks|Version| Filter |n-shot| Metric |Value | |Stderr|
|-----|-------|----------|-----:|-----------|-----:|---|-----:|
|gsm8k|Yaml |get-answer| 5|exact_match|0.6467|± |0.0132|
hf (pretrained=fblgit/LUNA-SOLARkrautLM-Instruct), gen_kwargs: (), limit: None, num_fewshot: 0, batch_size: auto (64)
| Tasks |Version|Filter|n-shot|Metric|Value | |Stderr|
|--------------|-------|------|-----:|------|-----:|---|-----:|
|truthfulqa_mc2|Yaml |none | 0|acc |0.7368|± |0.0149|
hf (pretrained=fblgit/LUNA-SOLARkrautLM-Instruct), gen_kwargs: (), limit: None, num_fewshot: 25, batch_size: auto (32)
| Tasks |Version|Filter|n-shot| Metric |Value| |Stderr|
|-------------|-------|------|-----:|--------|----:|---|-----:|
|arc_challenge|Yaml |none | 25|acc |0.692|± |0.0135|
| | |none | 25|acc_norm|0.715|± |0.0132|
hf (pretrained=fblgit/LUNA-SOLARkrautLM-Instruct), gen_kwargs: (), limit: None, num_fewshot: 0, batch_size: auto (64)
| Tasks |Version|Filter|n-shot|Metric| Value | |Stderr|
|-----------|-------|------|-----:|------|------:|---|-----:|
|paws_de |Yaml |none | 0|acc | 0.3965|± |0.0109|
|wmt16-en-de|Yaml |none | 0|bleu | 3.5784|± |0.1325|
| | |none | 0|ter |64.5707|± |0.4514|
| | |none | 0|chrf |45.7068|± |0.3861|
|xnli_de |Yaml |none | 0|acc | 0.4129|± |0.0099|
hf (pretrained=fblgit/LUNA-SOLARkrautLM-Instruct), gen_kwargs: (), limit: None, num_fewshot: 10, batch_size: auto (32)
| Tasks |Version|Filter|n-shot| Metric |Value | |Stderr|
|---------|-------|------|-----:|--------|-----:|---|-----:|
|hellaswag|Yaml |none | 10|acc |0.7131|± |0.0045|
| | |none | 10|acc_norm|0.8815|± |0.0032|
hf (pretrained=fblgit/LUNA-SOLARkrautLM-Instruct), gen_kwargs: (), limit: None, num_fewshot: 5, batch_size: auto (64)
| Tasks |Version|Filter|n-shot|Metric| Value | |Stderr|
|-----------|-------|------|-----:|------|------:|---|-----:|
|wmt16-de-en|Yaml |none | 5|bleu |14.9310|± |0.8014|
| | |none | 5|ter |46.3206|± |0.4087|
| | |none | 5|chrf |60.8637|± |0.4436|
|wmt16-en-de|Yaml |none | 5|bleu | 6.2016|± |0.2918|
| | |none | 5|ter |63.9997|± |0.4591|
| | |none | 5|chrf |51.1399|± |0.3978|
|xnli_de |Yaml |none | 5|acc | 0.4703|± |0.0100|
hf (pretrained=fblgit/LUNA-SOLARkrautLM-Instruct,dtype=float16), gen_kwargs: (), limit: None, num_fewshot: 5, batch_size: auto (16)
| Tasks |Version|Filter|n-shot|Metric|Value | |Stderr|
|---------------------------------------|-------|------|-----:|------|-----:|---|-----:|
|mmlu |N/A |none | 0|acc |0.6461|± |0.1215|
| - humanities |N/A |none | 5|acc |0.5960|± |0.1200|
| - formal_logic |Yaml |none | 5|acc |0.4683|± |0.0446|
| - high_school_european_history |Yaml |none | 5|acc |0.8121|± |0.0305|
| - high_school_us_history |Yaml |none | 5|acc |0.8480|± |0.0252|
| - high_school_world_history |Yaml |none | 5|acc |0.8312|± |0.0244|
| - international_law |Yaml |none | 5|acc |0.7851|± |0.0375|
| - jurisprudence |Yaml |none | 5|acc |0.7685|± |0.0408|
| - logical_fallacies |Yaml |none | 5|acc |0.7423|± |0.0344|
| - moral_disputes |Yaml |none | 5|acc |0.7283|± |0.0239|
| - moral_scenarios |Yaml |none | 5|acc |0.3899|± |0.0163|
| - philosophy |Yaml |none | 5|acc |0.7074|± |0.0258|
| - prehistory |Yaml |none | 5|acc |0.7716|± |0.0234|
| - professional_law |Yaml |none | 5|acc |0.4824|± |0.0128|
| - world_religions |Yaml |none | 5|acc |0.7661|± |0.0325|
| - other |N/A |none | 5|acc |0.7097|± |0.0900|
| - business_ethics |Yaml |none | 5|acc |0.7700|± |0.0423|
| - clinical_knowledge |Yaml |none | 5|acc |0.6792|± |0.0287|
| - college_medicine |Yaml |none | 5|acc |0.6647|± |0.0360|
| - global_facts |Yaml |none | 5|acc |0.3600|± |0.0482|
| - human_aging |Yaml |none | 5|acc |0.6861|± |0.0311|
| - management |Yaml |none | 5|acc |0.8350|± |0.0368|
| - marketing |Yaml |none | 5|acc |0.8504|± |0.0234|
| - medical_genetics |Yaml |none | 5|acc |0.6700|± |0.0473|
| - miscellaneous |Yaml |none | 5|acc |0.7893|± |0.0146|
| - nutrition |Yaml |none | 5|acc |0.7549|± |0.0246|
| - professional_accounting |Yaml |none | 5|acc |0.5213|± |0.0298|
| - professional_medicine |Yaml |none | 5|acc |0.7353|± |0.0268|
| - virology |Yaml |none | 5|acc |0.5783|± |0.0384|
| - social_sciences |N/A |none | 5|acc |0.7501|± |0.0684|
| - econometrics |Yaml |none | 5|acc |0.5175|± |0.0470|
| - high_school_geography |Yaml |none | 5|acc |0.8485|± |0.0255|
| - high_school_government_and_politics|Yaml |none | 5|acc |0.8912|± |0.0225|
| - high_school_macroeconomics |Yaml |none | 5|acc |0.6615|± |0.0240|
| - high_school_microeconomics |Yaml |none | 5|acc |0.7311|± |0.0288|
| - high_school_psychology |Yaml |none | 5|acc |0.8385|± |0.0158|
| - human_sexuality |Yaml |none | 5|acc |0.7023|± |0.0401|
| - professional_psychology |Yaml |none | 5|acc |0.6683|± |0.0190|
| - public_relations |Yaml |none | 5|acc |0.6909|± |0.0443|
| - security_studies |Yaml |none | 5|acc |0.7633|± |0.0272|
| - sociology |Yaml |none | 5|acc |0.8358|± |0.0262|
| - us_foreign_policy |Yaml |none | 5|acc |0.8800|± |0.0327|
| - stem |N/A |none | 5|acc |0.5569|± |0.1360|
| - abstract_algebra |Yaml |none | 5|acc |0.3800|± |0.0488|
| - anatomy |Yaml |none | 5|acc |0.6148|± |0.0420|
| - astronomy |Yaml |none | 5|acc |0.7237|± |0.0364|
| - college_biology |Yaml |none | 5|acc |0.7708|± |0.0351|
| - college_chemistry |Yaml |none | 5|acc |0.4600|± |0.0501|
| - college_computer_science |Yaml |none | 5|acc |0.5400|± |0.0501|
| - college_mathematics |Yaml |none | 5|acc |0.2700|± |0.0446|
| - college_physics |Yaml |none | 5|acc |0.3333|± |0.0469|
| - computer_security |Yaml |none | 5|acc |0.7300|± |0.0446|
| - conceptual_physics |Yaml |none | 5|acc |0.6213|± |0.0317|
| - electrical_engineering |Yaml |none | 5|acc |0.6276|± |0.0403|
| - elementary_mathematics |Yaml |none | 5|acc |0.4788|± |0.0257|
| - high_school_biology |Yaml |none | 5|acc |0.8065|± |0.0225|
| - high_school_chemistry |Yaml |none | 5|acc |0.5123|± |0.0352|
| - high_school_computer_science |Yaml |none | 5|acc |0.7000|± |0.0461|
| - high_school_mathematics |Yaml |none | 5|acc |0.3889|± |0.0297|
| - high_school_physics |Yaml |none | 5|acc |0.3576|± |0.0391|
| - high_school_statistics |Yaml |none | 5|acc |0.5926|± |0.0335|
| - machine_learning |Yaml |none | 5|acc |0.4554|± |0.0473|
| Groups |Version|Filter|n-shot|Metric|Value | |Stderr|
|------------------|-------|------|-----:|------|-----:|---|-----:|
|mmlu |N/A |none | 0|acc |0.6461|± |0.1215|
| - humanities |N/A |none | 5|acc |0.5960|± |0.1200|
| - other |N/A |none | 5|acc |0.7097|± |0.0900|
| - social_sciences|N/A |none | 5|acc |0.7501|± |0.0684|
| - stem |N/A |none | 5|acc |0.5569|± |0.1360|
```
### MT-Bench
```
########## Average ##########
score
model
gpt-4 8.990625
gpt-3.5-turbo 7.943750
claude-instant-v1 7.905660
claude-v1 7.900000
UNA-SOLAR-10.7B-Instruct-v1.0 7.521875
LUNA-SOLARkrautLM-Instruct 7.462500
vicuna-33b-v1.3 7.121875
wizardlm-30b 7.009375
Llama-2-70b-chat 6.856250
Llama-2-13b-chat 6.650000
guanaco-33b 6.528125
tulu-30b 6.434375
guanaco-65b 6.409375
oasst-sft-7-llama-30b 6.409375
palm-2-chat-bison-001 6.400000
mpt-30b-chat 6.393750
vicuna-13b-v1.3 6.387500
wizardlm-13b 6.353125
Llama-2-7b-chat 6.268750
vicuna-7b-v1.3 5.996875
baize-v2-13b 5.750000
nous-hermes-13b 5.553459
mpt-7b-chat 5.459119
gpt4all-13b-snoozy 5.452830
koala-13b 5.350000
mpt-30b-instruct 5.218750
falcon-40b-instruct 5.168750
h2ogpt-oasst-open-llama-13b 4.625000
alpaca-13b 4.531250
chatglm-6b 4.500000
oasst-sft-4-pythia-12b 4.318750
rwkv-4-raven-14b 3.984375
dolly-v2-12b 3.275000
fastchat-t5-3b 3.040625
stablelm-tuned-alpha-7b 2.753125
llama-13b 2.606250
```
## Disclaimer
We must inform users that despite our best efforts in data cleansing, the possibility of uncensored content slipping through cannot be entirely ruled out.
However, we cannot guarantee consistently appropriate behavior. Therefore, if you encounter any issues or come across inappropriate content, we kindly request that you inform us through the contact information provided.
Additionally, it is essential to understand that the licensing of these models does not constitute legal advice. We are not held responsible for the actions of third parties who utilize our models.
## Contact
If you are interested in customized LLMs for business applications, please get in contact with us via our website or contact us at [Dr. Daryoush Vaziri](mailto:[email protected]). We are also grateful for your feedback and suggestions.
## Collaborations
We are also keenly seeking support and investment for our startup, [VAGO Solutions](https://huggingface.co/VAGOsolutions), where we continuously advance the development of robust language models designed to address a diverse range of purposes and requirements. If the prospect of collaboratively navigating future challenges excites you, we warmly invite you to reach out to us.
[Juanako.AI](https://huggingface.co/fblgit) is also seeking support and investment for our startup, we also are open for collaborating with other labs to make awesome models like this one.
## Acknowledgement
Big Hug to [VAGO Solutions](https://huggingface.co/VAGOsolutions), we merely used our UNA transformers library on their code and dataset, nothing else. This won't be possible without them, thanks!
Many thanks to [argilla](https://huggingface.co/datasets/argilla) and [Huggingface](https://huggingface.co) for providing such valuable datasets to the Open-Source community. And of course a big thanks to [upstage](https://huggingface.co/upstage) for providing the open source community with their latest technology!
<!-- original-model-card end -->
| [
"TRANSLATION"
] | Non_BioNLP |
halee9/translation_en_ko | halee9 | text2text-generation | [
"transformers",
"tensorboard",
"safetensors",
"marian",
"text2text-generation",
"generated_from_trainer",
"base_model:Helsinki-NLP/opus-mt-ko-en",
"base_model:finetune:Helsinki-NLP/opus-mt-ko-en",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,710,610,256,000 | 2024-03-16T22:43:22 | 128 | 0 | ---
base_model: Helsinki-NLP/opus-mt-ko-en
license: apache-2.0
metrics:
- bleu
tags:
- generated_from_trainer
model-index:
- name: translation_en_ko
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# translation_en_ko
This model is a fine-tuned version of [Helsinki-NLP/opus-mt-ko-en](https://huggingface.co/Helsinki-NLP/opus-mt-ko-en) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 1.4074
- Bleu: 30.5108
- Gen Len: 42.414
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Bleu | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|
| 1.5644 | 1.0 | 7500 | 1.4721 | 29.3866 | 42.268 |
| 1.3933 | 2.0 | 15000 | 1.4074 | 30.5108 | 42.414 |
### Framework versions
- Transformers 4.38.2
- Pytorch 2.2.1+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2
| [
"TRANSLATION"
] | Non_BioNLP |
lamm-mit/Cephalo-Idefics-2-vision-10b-beta | lamm-mit | image-text-to-text | [
"transformers",
"safetensors",
"idefics2",
"image-text-to-text",
"nlp",
"code",
"vision",
"chemistry",
"engineering",
"biology",
"bio-inspired",
"text-generation-inference",
"materials science",
"conversational",
"multilingual",
"arxiv:2405.19076",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | 1,716,909,925,000 | 2024-05-30T10:34:41 | 12 | 0 | ---
language:
- multilingual
library_name: transformers
license: apache-2.0
pipeline_tag: image-text-to-text
tags:
- nlp
- code
- vision
- chemistry
- engineering
- biology
- bio-inspired
- text-generation-inference
- materials science
inference:
parameters:
temperature: 0.3
widget:
- messages:
- role: user
content: <|image_1|>Can you describe what you see in the image?
---
## Model Summary
Cephalo is a series of multimodal materials science focused vision large language models (V-LLMs) designed to integrate visual and linguistic data for advanced understanding and interaction in human-AI or multi-agent AI frameworks.
A novel aspect of Cephalo's development is the innovative dataset generation method. The extraction process employs advanced algorithms to accurately detect and separate images and their corresponding textual descriptions from complex PDF documents. It involves extracting images and captions from PDFs to create well-reasoned image-text pairs, utilizing large language models (LLMs) for natural language processing. These image-text pairs are then refined and validated through LLM-based NLP processing, ensuring high-quality and contextually relevant data for training.
Cephalo can interpret complex visual scenes and generating contextually accurate language descriptions and answer queries.
The model is developed to process diverse inputs, including images and text, facilitating a broad range of applications such as image captioning, visual question answering, and multimodal content generation. The architecture combines a vision encoder model and an autoregressive transformer to process complex natural language understanding.
Cephalo provides a robust framework for multimodal interaction and understanding, including the development of complex generative pipelines to create 2D and 3D renderings of material microstructures as input for additive manufacturing methods.
This version of Cephalo, lamm-mit/Cephalo-Idefics-2-vision-10b-beta, is based on a merged expansion of the https://huggingface.co/lamm-mit/Cephalo-Idefics-2-vision-8b-beta and the HuggingFaceM4/idefics2-8b-chatty model. This method allows us to increase the depth of the model and focus on learning more complex representations and associations in deeper layers of the network.
The lamm-mit/Cephalo-Idefics-2-vision-10b-beta model is trained for two epochs, while the lamm-mit/Cephalo-Idefics-2-vision-10b-alpha version was trained for one epoch.
The model was trained in several stages:
**Step 1**: Train https://huggingface.co/lamm-mit/Cephalo-Idefics-2-vision-8b-beta by fine-tuning the HuggingFaceM4/idefics2-8b-chatty model.
**Step 2**: Combine the https://huggingface.co/lamm-mit/Cephalo-Idefics-2-vision-8b-beta decoder with the last 8 layers of the HuggingFaceM4/idefics2-8b-chatty decoder.
**Step 3**: Fine-tune the merged model, which now has 40 decoder layers and a total of 10b parameters.
The model was trained on a combination of scientific text-image data extracted from Wikipedia and scientific papers. For further details on the base model, see: https://huggingface.co/HuggingFaceM4/idefics2-8b-chatty. More details about technical aspects of the model, training and example applications to materials science problems are provided in the paper (reference at the bottom).
### Chat Format
The lamm-mit/Cephalo-Idefics-2-vision-10b-beta model is suitable for one or more image inputs, wih prompts using the chat format as follows:
```raw
User: You carefully study the image, and respond accurately, but succinctly. Think step-by-step.
<image>What is shown in this image, and what is the relevance for materials design? Include a discussion of multi-agent AI.<end_of_utterance>
Assistant:
```
where the model generates the text after `Assistant:` . For multi-turn conversations, the prompt should be formatted as follows:
```raw
User: You carefully study the image, and respond accurately, but succinctly. Think step-by-step.
<image>What is shown in this image, and what is the relevance for materials design? Include a discussion of multi-agent AI.<end_of_utterance>
Assistant: The image depicts ants climbing a vertical surface using their legs and claws. This behavior is observed in nature and can inspire the design of multi-agent AI systems that mimic the coordinated movement of these insects. The relevance lies in the potential application of such systems in robotics and materials science, where efficient and adaptive movement is crucial.<end_of_utterance>
User: How could this be used to design a fracture resistant material?<end_of_utterance>
Assistant:
```
If you need to manually set the chat template:
```
IDEFICS2_CHAT_TEMPLATE = "{% for message in messages %}{{message['role'].capitalize()}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>\n{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}"
```
### Sample inference code
This code snippets show how to get quickly started on a GPU:
```python
from PIL import Image
import requests
DEVICE='cuda:0'
from transformers import AutoProcessor, Idefics2ForConditionalGeneration
from tqdm.notebook import tqdm
model_id='lamm-mit/Cephalo-Idefics-2-vision-10b-beta'
model = Idefics2ForConditionalGeneration.from_pretrained( model_id,
torch_dtype=torch.bfloat16, #if your GPU allows
_attn_implementation="flash_attention_2", #make sure Flash Attention 2 is installed
trust_remote_code=True,
).to (DEVICE)
processor = AutoProcessor.from_pretrained(
f"{model_id}",
do_image_splitting=True
)
```
See section towards the end for more comments on model optimization, including quantization.
If you need to manually set the chat template:
```python
IDEFICS2_CHAT_TEMPLATE = "{% for message in messages %}{{message['role'].capitalize()}}{% if message['content'][0]['type'] == 'image' %}{{':'}}{% else %}{{': '}}{% endif %}{% for line in message['content'] %}{% if line['type'] == 'text' %}{{line['text']}}{% elif line['type'] == 'image' %}{{ '<image>' }}{% endif %}{% endfor %}<end_of_utterance>\n{% endfor %}{% if add_generation_prompt %}{{ 'Assistant:' }}{% endif %}"
tokenizer = AutoTokenizer.from_pretrained(base_model_id, use_fast=True)
tokenizer.chat_template = IDEFICS2_CHAT_TEMPLATE
processor.tokenizer = tokenizer
```
Simple inference example:
```
from transformers.image_utils import load_image
image = load_image("https://d2r55xnwy6nx47.cloudfront.net/uploads/2018/02/Ants_Lede1300.jpg")
# Create inputs
messages = [
{
"role": "user",
"content": [
{"type": "image"},
{"type": "text", "text": "What is shown in this image, and what is the relevance for materials design? Include a discussion of multi-agent AI."},
]
},
]
prompt = processor.apply_chat_template(messages, add_generation_prompt=True)
# Get inputs using the processor
inputs = processor(text=prompt, images=[image], return_tensors="pt")
inputs = {k: v.to(DEVICE) for k, v in inputs.items()}
# Generate
generated_ids = model.generate(**inputs, max_new_tokens=500)
generated_texts = processor.batch_decode(generated_ids, skip_special_tokens=True)
print(generated_texts)
```
Next we provide a convenience function for inference. This function takes the model, processor, question, and images, along with messages and images objects for repeated chat-like interactions with the model.
```python
def ask_about_image (model, processor, question,
images_input=[],
verbatim=False,
temperature=0.1,
show_image=False,
system="You are a biomaterials scientist who responds accurately. ",
init_instr = "",
show_conversation=True,
max_new_tokens=256,
messages=[],
images=[],
use_Markdown=False,
):
query = question
images_input=ensure_list(images_input)
if len (images)==0:
if len (images_input)>0:
for image in tqdm (images_input) :
if is_url(image):
image= load_image(image)
images.append (image)
if show_image:
display ( image )
if len (messages)==0:
base_message = {
"role": "user",
"content": [
{"type": "text", "text": system + init_instr},
# Image messages will be added dynamically here
{"type": "text", "text": query}
]
}
# Ensure the images_input is a list
images_input = ensure_list(images_input)
# Add image messages dynamically
image_messages = [{"type": "image"} for _ in images_input]
base_message["content"][1:1] = image_messages # Insert image messages before the last text message
# Append the constructed message to messages list
messages.append(base_message)
else:
messages.append (
{
"role": "user",
"content": [
{"type": "text", "text": query
}
]
}
)
if verbatim:
print (messages)
text = processor.apply_chat_template(messages, add_generation_prompt=True)
inputs = processor(text=[text.strip()], images=images, return_tensors="pt", padding=True).to(DEVICE)
generated_ids = model.generate(**inputs, max_new_tokens=max_new_tokens, temperature=temperature, do_sample=True)
generated_texts = processor.batch_decode(generated_ids[:, inputs["input_ids"].size(1):], skip_special_tokens=True)
messages.append (
{
"role": "assistant",
"content": [ {"type": "text", "text": generated_texts[0]}, ]
}
)
formatted_conversation = format_conversation(messages, images)
# Display the formatted conversation, e.g. in Jupyter Notebook
if show_conversation:
if use_Markdown:
display(Markdown(formatted_conversation))
else:
display(HTML(formatted_conversation))
return generated_texts, messages, images
question = "What is shown in this image, and what is the relevance for materials design? Include a discussion of multi-agent AI."
url1 = "https://d2r55xnwy6nx47.cloudfront.net/uploads/2018/02/Ants_Lede1300.jpg"
response, messages,images= ask_about_image ( model, processor, question,
images_input=[url1,],
temperature=0.1,
system= '', init_instr='You carefully study the image and provide detailed answers. Think step-by-step.\n\n',
show_conversation=True,
max_new_tokens=512, messages=[], images=[])
```
Sample output:
<small>Image by [Vaishakh Manohar](https://www.quantamagazine.org/the-simple-algorithm-that-ants-use-to-build-bridges-20180226/)</small>
<pre style="white-space: pre-wrap;">
The image shows a group of ants moving in coordinated patterns on a surface. This illustrates the concept of multi-agent AI, which involves the study and simulation of complex systems involving multiple agents (in this case, ants) interacting with each other and their environment.
The relevance for materials design is in understanding how these natural systems exhibit emergent behaviors such as self-organization, which can inspire the development of new materials and systems that mimic these natural processes. By studying the movement patterns of ants, researchers can gain insights into how to design materials that exhibit similar emergent properties, leading to improved performance in various applications.
Multi-agent AI involves creating models that describe the interactions between individual agents and their environment, allowing for the simulation of complex systems with multiple interacting components. This approach can be applied to various fields, including materials science, where understanding emergent behaviors at the microscopic level can lead to the design of new materials with enhanced properties.
</pre>
## Dataset generation
The schematic below shows a visualization of the approach to generate datasets for training the vision model. The extraction process employs advanced algorithms to accurately detect and separate images and their corresponding textual descriptions from complex PDF documents. It involves extracting images and captions from PDFs to create well-reasoned image-text pairs, utilizing large language models (LLMs) for natural language processing. These image-text pairs are then refined and validated through LLM-based NLP processing, ensuring high-quality and contextually relevant data for training.
The image below shows reproductions of two representative pages of the scientific article (here, Spivak, Buehler, et al., 2011), and how they are used to extract visual scientific data for training the Cephalo model.
# Further model optimizations
If your GPU allows, load and run inference in half precision (`torch.float16` or `torch.bfloat16`).
```diff
model = AutoModelForVision2Seq.from_pretrained(
"lamm-mit/Cephalo-Idefics-2-vision-10b-beta",
+ torch_dtype=torch.float16,
).to(DEVICE)
```
**Vision encoder efficiency**
Given the high resolution supported, the vision part of the model can be memory hungry depending on your configuration. If you are GPU-memory-constrained, you can:
- **deactivate the image splitting.** To do so, add `do_image_splitting=False` when initializing the processor (`AutoProcessor.from_pretrained`). There are no changes required on the model side. Note that only the sft model has been trained with image splitting.
- **decrease the maximum image resolution.** To do so, add `size= {"longest_edge": 448, "shortest_edge": 378}` when initializing the processor (`AutoProcessor.from_pretrained`). In particular, the `longest_edge` value can be adapted to fit the need (the default value is `980`). We recommend using values that are multiples of 14. There are no changes required on the model side.
`do_image_splitting=True` is especially needed to boost performance on complex tasks where a very large image is used as input. The model was fine-tuned with image splitting turned on. For simple tasks, this argument can be safely set to `False`.
**Using Flash-attention 2 to speed up generation**
<details><summary>Click to expand.</summary>
Mke sure to install `flash-attn`. Refer to the [original repository of Flash Attention](https://github.com/Dao-AILab/flash-attention) for the package installation. Simply change the snippet above with:
```diff
model = AutoModelForVision2Seq.from_pretrained(
"lamm-mit/Cephalo-Idefics-2-vision-10b-beta",
+ torch_dtype=torch.bfloat16,
+ _attn_implementation="flash_attention_2",
).to(DEVICE)
```
</details>
**4 bit quantization with bitsandbytes**
<details><summary>Click to expand.</summary>
It is possible to load Cephalo-Idefics-2-vision-10b-beta in 4bits with `bitsandbytes`. Make sure that you have `accelerate` and `bitsandbytes` installed.
```diff
+ from transformers import BitsAndBytesConfig
quantization_config = BitsAndBytesConfig(
load_in_4bit=True,
bnb_4bit_quant_type="nf4",
bnb_4bit_use_double_quant=True,
bnb_4bit_compute_dtype=torch.bfloat16
)
model = AutoModelForVision2Seq.from_pretrained(
"lamm-mit/Cephalo-Idefics-2-vision-10b-beta",
+ torch_dtype=torch.bfloat16,
+ quantization_config=quantization_config,
).to(DEVICE)
```
</details>
## Citation
Please cite as:
```bibtex
@article{Buehler_Cephalo_2024,
title={Cephalo: Multi-Modal Vision-Language Models for Bio-Inspired Materials Analysis and Design},
author={Markus J. Buehler},
journal={arXiv preprint arXiv:2405.19076},
year={2024}
}
``` | [
"QUESTION_ANSWERING"
] | Non_BioNLP |
gauravkoradiya/T5-Finetuned-Summarization-DialogueDataset | gauravkoradiya | summarization | [
"transformers",
"pytorch",
"t5",
"text2text-generation",
"summarization",
"en",
"dataset:knkarthick/dialogsum",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,681,607,546,000 | 2023-04-16T01:24:14 | 151 | 1 | ---
datasets:
- knkarthick/dialogsum
language:
- en
library_name: transformers
license: apache-2.0
metrics:
- bleu
- rouge
pipeline_tag: summarization
---
| [
"SUMMARIZATION"
] | Non_BioNLP |
MaLA-LM/lucky52-bloom-7b1-no-5 | MaLA-LM | text-generation | [
"transformers",
"pytorch",
"bloom",
"text-generation",
"generation",
"question answering",
"instruction tuning",
"multilingual",
"dataset:MBZUAI/Bactrian-X",
"arxiv:2404.04850",
"license:cc-by-nc-4.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,712,217,803,000 | 2024-12-10T09:07:41 | 14 | 0 | ---
datasets:
- MBZUAI/Bactrian-X
language:
- multilingual
library_name: transformers
license: cc-by-nc-4.0
pipeline_tag: text-generation
tags:
- generation
- question answering
- instruction tuning
---
### Model Description
This HF repository hosts instruction fine-tuned multilingual BLOOM model using the parallel instruction dataset called Bactrain-X in 52 languages.
We progressively add a language during instruction fine-tuning at each time, and train 52 models in total. Then, we evaluate those models in three multilingual benchmarks.
Please refer to [our paper](https://arxiv.org/abs/2404.04850) for more details.
* Base model: [BLOOM 7B1](https://huggingface.co/bigscience/bloom-7b1)
* Instruction languages: English, Chinese, Afrikaans, Arabic, Azerbaijani
* Instruction language codes: en, zh, af, ar, az
* Training method: full-parameter fine-tuning.
### Usage
The model checkpoint should be loaded using `transformers` library.
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("MaLA-LM/lucky52-bloom-7b1-no-5")
model = AutoModelForCausalLM.from_pretrained("MaLA-LM/lucky52-bloom-7b1-no-5")
```
### Citation
```
@inproceedings{ji2025lucky52,
title={How Many Languages Make Good Multilingual Instruction Tuning? A Case Study on BLOOM},
author={Shaoxiong Ji and Pinzhen Chen},
year={2025},
booktitle={Proceedings of COLING},
url={https://arxiv.org/abs/2404.04850},
}
```
| [
"QUESTION_ANSWERING"
] | Non_BioNLP |
RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf | RichardErkhov | null | [
"gguf",
"endpoints_compatible",
"region:us"
] | 1,721,885,085,000 | 2024-07-25T11:07:58 | 26 | 0 | ---
{}
---
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
airoboros-l2-13b-3.0 - GGUF
- Model creator: https://huggingface.co/jondurbin/
- Original model: https://huggingface.co/jondurbin/airoboros-l2-13b-3.0/
| Name | Quant method | Size |
| ---- | ---- | ---- |
| [airoboros-l2-13b-3.0.Q2_K.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q2_K.gguf) | Q2_K | 4.52GB |
| [airoboros-l2-13b-3.0.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.IQ3_XS.gguf) | IQ3_XS | 4.99GB |
| [airoboros-l2-13b-3.0.IQ3_S.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.IQ3_S.gguf) | IQ3_S | 5.27GB |
| [airoboros-l2-13b-3.0.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q3_K_S.gguf) | Q3_K_S | 5.27GB |
| [airoboros-l2-13b-3.0.IQ3_M.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.IQ3_M.gguf) | IQ3_M | 5.57GB |
| [airoboros-l2-13b-3.0.Q3_K.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q3_K.gguf) | Q3_K | 5.9GB |
| [airoboros-l2-13b-3.0.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q3_K_M.gguf) | Q3_K_M | 5.9GB |
| [airoboros-l2-13b-3.0.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q3_K_L.gguf) | Q3_K_L | 6.45GB |
| [airoboros-l2-13b-3.0.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.IQ4_XS.gguf) | IQ4_XS | 6.54GB |
| [airoboros-l2-13b-3.0.Q4_0.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q4_0.gguf) | Q4_0 | 6.86GB |
| [airoboros-l2-13b-3.0.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.IQ4_NL.gguf) | IQ4_NL | 6.9GB |
| [airoboros-l2-13b-3.0.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q4_K_S.gguf) | Q4_K_S | 6.91GB |
| [airoboros-l2-13b-3.0.Q4_K.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q4_K.gguf) | Q4_K | 7.33GB |
| [airoboros-l2-13b-3.0.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q4_K_M.gguf) | Q4_K_M | 7.33GB |
| [airoboros-l2-13b-3.0.Q4_1.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q4_1.gguf) | Q4_1 | 7.61GB |
| [airoboros-l2-13b-3.0.Q5_0.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q5_0.gguf) | Q5_0 | 8.36GB |
| [airoboros-l2-13b-3.0.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q5_K_S.gguf) | Q5_K_S | 8.36GB |
| [airoboros-l2-13b-3.0.Q5_K.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q5_K.gguf) | Q5_K | 8.6GB |
| [airoboros-l2-13b-3.0.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q5_K_M.gguf) | Q5_K_M | 8.6GB |
| [airoboros-l2-13b-3.0.Q5_1.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q5_1.gguf) | Q5_1 | 9.1GB |
| [airoboros-l2-13b-3.0.Q6_K.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q6_K.gguf) | Q6_K | 9.95GB |
| [airoboros-l2-13b-3.0.Q8_0.gguf](https://huggingface.co/RichardErkhov/jondurbin_-_airoboros-l2-13b-3.0-gguf/blob/main/airoboros-l2-13b-3.0.Q8_0.gguf) | Q8_0 | 12.88GB |
Original model description:
---
license: llama2
datasets:
- jondurbin/airoboros-3.0
---
### Overview
Another experimental model, using mostly sythetic data generated by [airoboros](https://github.com/jondurbin/airoboros)
#### IMPORTANT NOTE - llama-2 chat format!!!
The airobors 3.0 models have been updated to use llama-2 chat format, rather than vicuna style user/assistant!
This is a breaking change, although most inference systems support llama-2 chat templating.
See prompt format below for all of the details.
#### Highlights
A model built on [airoboros-3.0 dataset](https://hf.co/datasets/jondurbin/airoboros-3.0), which is a continuation of [airoboros-2.2.1 dataset](https://hf.co/datasets/jondurbin/airoboros-2.2.1) with the following extras:
* MathJSON - this is not yet part of the code directly, but can be easily parsed and executed, as a way to perform deterministic calculations. LLMs suck at math, so this does problem formulation and let's the work be done by traditional operations.
* Human-generated RP multi-turn conversation - provided by an anon, this is a human created RP dataset, meant to assist with multi-turn coherency and chat experience.
This is a fairly general purpose model, but focuses heavily on instruction following, rather than casual chat/roleplay.
Huge thank you to the folks over at [a16z](https://a16z.com/) for sponsoring the costs associated with building models and associated tools!
### Prompt format
The prompt format is llama-2 chat.
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt} [/INST]
```
For multi-turn, the prompt format is as follows:
```
[INST] <<SYS>>
You are a help, unbiased, uncensored assistant.
<</SYS>
{prompt 0} [/INST] {response 0} </s><s>[INST] {prompt 1} [/INST] {response 1} </s><s>...[INST] {prompt N} [/INST]
```
Here is a simple python script that can help generate the prompt:
```python
def get_chat_prompt(message, chat_history, system_prompt):
texts = [f'[INST] <<SYS>>\n{system_prompt}\n<</SYS>>\n\n']
do_strip = False
for user_input, response in chat_history:
user_input = user_input.strip() if do_strip else user_input
do_strip = True
texts.append(f'{user_input} [/INST] {response.strip()} </s><s>[INST] ')
message = message.strip() if do_strip else message
texts.append(f'{message} [/INST] ')
return ''.join(texts)
def get_prompt(item):
"""
`item` is expected to be in chatml format, e.g.
{
"id": 1,
"conversations": [
{
"role": "system",
"content": "You are a helpful, unbiased, uncensored assistant."
},
{
"role": "user",
"content": "Hi!"
},
{
"role": "assistant",
"content": "Hello, how are you today?"
},
{
"role": "user",
"content": "Just peachy, and you?"
}
]
}
"""
system_prompt = share_gpt_item['conversations'][0]['value']
offset = 1
if share_gpt_item['conversations'][0]['role'] != 'system':
system_prompt = 'You are an unbiased, uncensored, helpful assistant.'
offset = 0
chat_history = [
(share_gpt_item['conversations'][idx]['content'], share_gpt_item['conversations'][idx + 1]['content'])
for idx in range(offset, len(item['conversations']) - 1, 2)
]
message = item['conversations'][-1]['content']
return get_chat_prompt(message, chat_history, system_prompt)
```
### Fine tuning info
https://wandb.ai/jondurbin/airoboros-l2-13b-3.0/runs/97utxyk9?workspace=user-jondurbin
### Helpful usage tips
#### MathJSON
Ask mathematical questions, prefixed with "Create a MathJSON solution to the following:", and you should get output with a formula that can be executed by https://cortexjs.io/compute-engine/
I also made a (really quite hacky, incomplete, semi-functional) python version that works for most expressions, see [mathjson.py](https://github.com/jondurbin/airoboros/blob/mathjson/airoboros/mathjson.py)
__Use a very low temperature!__
Example:
```
[INST] <<SYS>>
You are an assistant with vast knowledge in all things math.
<</SYS>
Create a MathJSON solution to the following: Calculate the area of a circle with radius of 17.2456cm. Include your reasoning. [/INST]
```
Output:
```
The area (A) of a circle is given by the formula A = πr², where r is the radius of the circle. Since the radius is given in centimeters, we need to convert it to meters first. We know that 1 meter equals 100 centimeters. Therefore, we can represent the radius as: r = 17.2456 cm * 100/1.
Once we have the radius in meters, we can calculate the area of the circle using the formula A = πr².
Solution as MathJSON:
<mathjson>
[
"Multiply",
"Pi",
[
"Power",
17.2456,
2
]
]
</mathjson>
```
You can then validate the JSON between `<mathjson>` and `</mathjson>`, then pass the parsed JSON to compute-engine JS or the `evaluate` function in mathjson.py to calculate the response.
#### Context obedient question answering
By obedient, I mean the model was trained to ignore what it thinks it knows, and uses the context to answer the question. The model was also tuned to limit the values to the provided context as much as possible to reduce hallucinations.
The format for a closed-context prompt is as follows:
```
BEGININPUT
BEGINCONTEXT
[key0: value0]
[key1: value1]
... other metdata ...
ENDCONTEXT
[insert your text blocks here]
ENDINPUT
[add as many other blocks, in the exact same format]
BEGININSTRUCTION
[insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.]
ENDINSTRUCTION
```
It's also helpful to add "Don't make up answers if you don't know." to your instruction block to make sure if the context is completely unrelated it doesn't make something up.
*The __only__ prompts that need this closed context formating are closed-context instructions. Normal questions/instructions do not!*
I know it's a bit verbose and annoying, but after much trial and error, using these explicit delimiters helps the model understand where to find the responses and how to associate specific sources with it.
- `BEGININPUT` - denotes a new input block
- `BEGINCONTEXT` - denotes the block of context (metadata key/value pairs) to associate with the current input block
- `ENDCONTEXT` - denotes the end of the metadata block for the current input
- [text] - Insert whatever text you want for the input block, as many paragraphs as can fit in the context.
- `ENDINPUT` - denotes the end of the current input block
- [repeat as many input blocks in this format as you want]
- `BEGININSTRUCTION` - denotes the start of the list (or one) instruction(s) to respond to for all of the input blocks above.
- [instruction(s)]
- `ENDINSTRUCTION` - denotes the end of instruction set
It sometimes works without `ENDINSTRUCTION`, but by explicitly including that in the prompt, the model better understands that all of the instructions in the block should be responded to.
__Use a very low temperature!__
Here's a trivial, but important example to prove the point:
```
BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
```
And the response:
```
Blueberries are now green.
Source:
date: 2021-01-01
url: https://web.site/123
```
#### Summarization
500 samples have been included from [this dataset](https://huggingface.co/datasets/mattpscott/airoboros-summarization), using the same format as contextual question answering, for example:
```
BEGININPUT
{text to summarize}
ENDINPUT
BEGININSTRUCTION
Summarize the input in around 130 words.
ENDINSTRUCTION
```
#### Getting longer responses
You can use a few techniques to get longer responses.
Detailed prompts, with explicit instruction for word count:
```
Please compose a narrative set in the heart of an ancient library, steeped in the scent of old parchment and ink. The protagonist should be a young scholar who is dedicated to studying the art of storytelling and its evolution throughout history. In her pursuit of knowledge, she stumbles upon a forgotten tome that seems to possess an unusual aura. This book has the ability to bring stories to life, literally manifesting characters and scenarios from within its pages into reality.
The main character must navigate through various epochs of storytelling - from oral traditions of tribal societies, through medieval minstrels' tales, to modern-day digital narratives - as they come alive around her. Each era presents its unique challenges and lessons about the power and impact of stories on human civilization.
One such character could be a sentient quill pen, who was once used by renowned authors of yesteryears and now holds their wisdom and experiences. It becomes her mentor, guiding her through this journey with witty remarks and insightful commentary.
Ensure that your tale encapsulates the thrill of adventure, the beauty of learning, and the profound connection between humans and their stories. All characters involved should be non-human entities. Feel free to explore creative liberties but maintain the mentioned elements.
Your response should be approximately 2300 words.
```
Or, a simpler example:
```
Please create a long, detailed story about a dragon in an old growth forest who, for some reason, begins speaking the words of the source code of linux.
```
There are a few examples of next chapter completion as well, e.g.:
```
Write the next chapter of a historical fiction novel set in Paris during the 20th century.
Here's a summary of the previous chapter:
In the vibrant city of Paris, amid the tumultuous changes of the 20th century, our protagonist Margot, an aspiring fashion designer, has just secured an apprenticeship at a prestigious couture house. She meets Lucien, a charming journalist who covers the fashion industry. Together they navigate the ever-changing world of fashion and society, uncovering secrets that reveal the intricate links between style, politics, and culture. As the chapter concludes, they decide to delve deeper into the hidden corners of the fashion world to unravel its mysteries.
Requirements for the next chapter:
1. Character Development of Margot and Lucien:
- Margot's Evolution: Unfold more about Margot's past, her dreams of revolutionizing fashion, and her struggle to establish herself in a male-dominated industry. Illustrate her growing expertise, innovative ideas, and increasing dependence on Lucien.
- Lucien's Complexity: Introduce uncertainties surrounding Lucien's background and real motives. Increase suspense by suggesting undisclosed information he possesses, while also highlighting his wit and perceptiveness.
2. Exploration of Paris and the Couture House:
- Paris: Elaborate their journey through the bustling streets of Paris, including encounters with iconic figures, social unrest, and relics from different eras of French history.
- The Couture House: Expand on the grandeur of the couture house they work in, filled with artistic masterpieces, intense competition, and cryptic notes hinting at a scandalous past.
3. Emergence of the Subplot: The Lost Collection:
- Discovery: Have Margot and Lucien stumble upon a secret vault containing a lost collection designed before World War II, raising new questions about the previous owner and the influence of war on fashion.
- Revelation: Capture their shock as they realize the designs were plagiarized, the potential repercussions, and the opportunities it presents for Margot's career.
- Twist: End with a twist that suggests there are other stolen collections across Paris, setting up their new mission.
Your response should be approximately 650 words.
```
#### Coding
You can ask for fairly complex coding instructions with multiple criteria, e.g.:
```
Create a python application with the following requirements:
- Asyncio FastAPI webserver
- ping endpoint that returns the current date in JSON format
- file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate
```
Or inline criteria:
```
Write a multi-threaded TCP server in C that accepts a "GET [key]" input and "SET [key] [value]" input, and uses a binary tree to get and store the input values.
```
You can also optionally add a single space and "PLAINFORMAT" at the end of your prompt to avoid backticks, explanations, etc. and just print the code, e.g.:
```
Write a websocket application in node.js. PLAINFORMAT
```
#### Agent/function calling
The dataset includes many examples of function/args generation based on input criteria. This is somewhat similar to the OpenAI function calling, but the output is either JSON or YAML.
Example prompt:
```
As an AI assistant, please select the most suitable function and parameters from the list of available functions below, based on the user's input. Provide your response in JSON format.
Input: I want to know how many times 'Python' is mentioned in my text file.
Available functions:
file_analytics:
description: This tool performs various operations on a text file.
params:
action: The operation we want to perform on the data, such as "count_occurrences", "find_line", etc.
filters:
keyword: The word or phrase we want to search for.
```
Response:
```json
{
"function": "file_analytics",
"params": {
"action": "count_occurrences",
"filters": {
"keyword": "Python"
}
}
}
```
#### Chain-of-thought
You can ask for several possible responses to a given problem, with a ranking and final answer selection.
Example prompt:
```
A group of people decide to play a round-robin tournament where every player plays every other player exactly once. If a total of 45 games were played, how many players participated in the tournament? Offer several potential answers, rank them according to mathematical logic, and pick the most feasible one.
```
Example response:
```
Solution 1:
In a round-robin tournament, each player plays every other player exactly once. This means that the total number of games played is equal to the combination of all players taken two at a time (since a game involves two players).
The formula for combinations is nC2 = n(n-1)/2, where n is the total number of players. We know that nC2 equals 45 (the total number of games), so we can set up the equation as follows:
n(n-1)/2 = 45
Solving this equation gives us n=10.
Final answer: There were 10 players in the tournament.
Solution 2:
Another way to approach this problem is by using the concept of triangular numbers. Triangular numbers are a sequence of numbers where each number can be represented as a triangle with dots. The nth triangular number is given by the sum of the first n natural numbers.
If we consider each game as a dot and each player as a side of the triangle, then the total number of games (45) would correspond to the 9th triangular number because 1+2+3+4+5+6+7+8+9=45. However, since each side of the triangle represents a player, and there's one more player than sides in our model (because the last player has no one left to play against), we need to add one more to get the total number of players.
Final answer: There were 10 players in the tournament.
Solution 3:
We could also use trial and error to solve this problem. Starting from a small number of players, we calculate how many games would be played in a round-robin tournament until we reach 45 games. For example, with 2 players, only 1 game is played; with 3 players, 3 games are played; with 4 players, 6 games are played, and so on. Continuing this process, we find that with 10 players, 45 games are played.
Final answer: There were 10 players in the tournament.
Ranking of solutions from best to worst:
1. Solution 1: This solution uses direct mathematical reasoning based on the properties of a round-robin tournament and solves an equation to find the answer.
2. Solution 2: This solution uses the concept of triangular numbers which requires additional steps and knowledge beyond the immediate problem context, but still arrives at the correct answer.
3. Solution 3: This solution relies on trial and error, which is less efficient and reliable than the other methods, but it eventually gets to the right answer.
Best and final answer: There were 10 players in the tournament.
```
#### reWOO style execution planning
The model now supports execution planning for complex instructions that would require making use of several tools. The output is just the plan, you must implement a mechanism to parse the output and actually call the functions!
Example prompt:
```
Please construct a systematic plan to generate an optimal response to the user instruction, utilizing a set of provided tools. Each plan will correspond to an evidence value, which will be the output of one of the available functions given an input string
that could be the user's question, one or more prior evidence values, or a combination of both.
Here are the tools available to be called:
DuckDuckGo[input]: Tool that enables the user to search for information using the DuckDuckGo search engine. This tool is helpful in locating a suitable list of sites that may contain the answer to the user's query. It does not directly answer the question but finds a list of sites that might have the answer.
TextScraper[input]: Load one or more websites from the input string containing newline delimited links, where input is one or more links, and produces plain text output containing the content of the links.
HyperlinkExtractor[input]: Extract links from plain text and produces a plain text, newline delimited response of links.
KnowledgeModel[input]: Question answering language model, particularly useful in answering questions based on an input passage of text. The input must be a text question that references an :evidence[n]: variable, e.g. What color is the cat, given :evidence1:?
The output should be in format:
Plan: [first action to take based on input question]
:evidence0: = [function to call with input parameter]
Plan: [next action to take, based on result of :evidence0:]
:evidence1: = [next function to call and input parameter, which may include reference to previous evidence, e.g. "Given context :evidence0"]
...
Answer: [:evidence[n]: containing the final answer.]
Question: Who won the Nobel Prize in Literature in 2020 and what is their most famous work?
```
Response:
```
Plan: Begin by conducting a web search to find out who won the Nobel Prize in Literature in 2020.
:evidence0: = DuckDuckGo[Nobel Prize in Literature 2020 winner]
Plan: Utilize the KnowledgeModel to interpret the search results and find the name of the laureate.
:evidence1: = KnowledgeModel[Who won the Nobel Prize in Literature in 2020, given :evidence0:?]
Plan: Conduct another web search to find the most famous work of the identified laureate.
:evidence2: = DuckDuckGo[Most famous work of :evidence1:]
Plan: Extract the relevant links from the DuckDuckGo search results for a more focused search.
:evidence3: = HyperlinkExtractor[:evidence2:]
Plan: Use the TextScraper tool to extract information from the relevant links.
:evidence4: = TextScraper[:evidence3:]
Plan: Finally, utilize the KnowledgeModel to identify and summarize the most famous work of the laureate from the extracted information.
:evidence5: = KnowledgeModel[What is the most famous work of :evidence1:, given :evidence4:?]
Answer: :evidence5:
```
For this to be useful, you'd have to parse the output plan text, and implement/call each of the functions. This is just pseudo-code, completely untested off the top of my head, and obviously would requiring full implementation + hardening:
```python
import re
import requests
def inject_context(input_text, **context):
for ref in set(re.findall(r"(:evidence[0-9]+:)", input_text, re.I)):
input_text = input_text.replace(ref, context.get(ref, ""))
return input_text
def duckduckgo(input_text, **context):
search_string = inject_context(input_text, **context)
... search via duck duck go using search_string
... return text content
def link_extractor(input_text, **context):
input_text = inject_context(input_text, **context)
return "\n".join(list(set(re.findall(r"(https?://[^\s]+?\.?)", input_text, re.I))))
def scrape(input_text, **context):
input_text = inject_context(input_text, **context)
text = []
for link in input_text.splitlines():
text.append(requests.get(link).text)
return "\n".join(text)
def infer(input_text, **context)
prompt = inject_context(input_text, **context)
... call model with prompt, return output
def parse_plan(plan):
method_map = {
"DuckDuckGo": duckduckgo,
"HyperlinkExtractor": link_extractor,
"KnowledgeModel": infer,
"TextScraper": scrape,
}
context = {}
for line in plan.strip().splitlines():
if line.startswith("Plan:"):
print(line)
continue
parts = re.match("^(:evidence[0-9]+:)\s*=\s*([^\[]+])(\[.*\])\s$", line, re.I)
if not parts:
if line.startswith("Answer: "):
return context.get(line.split(" ")[-1].strip(), "Answer couldn't be generated...")
raise RuntimeError("bad format: " + line)
context[parts.group(1)] = method_map[parts.group(2)](parts.group(3), **context)
```
### Contribute
If you're interested in new functionality, particularly a new "instructor" type to generate a specific type of training data,
take a look at the dataset generation tool repo: https://github.com/jondurbin/airoboros and either make a PR or open an issue with details.
To help me with the OpenAI/compute costs:
- https://bmc.link/jondurbin
- ETH 0xce914eAFC2fe52FdceE59565Dd92c06f776fcb11
- BTC bc1qdwuth4vlg8x37ggntlxu5cjfwgmdy5zaa7pswf
### Licence and usage restrictions
The airoboros 3.0 models are built on top of multiple base models, each with their own license/restrictions.
The models with `-l2` in the name have a custom Meta license:
- See the [meta-license/LICENSE.txt](meta-license/LICENSE.txt) file attached for the original license provided by Meta.
- See also [meta-license/USE_POLICY.md](meta-license/USE_POLICY.md) and [meta-license/Responsible-Use-Guide.pdf](meta-license/Responsible-Use-Guide.pdf), also provided by Meta.
The models with `-m-` are mistral-7b (apache 2.0)
The model with `-3b` uses Stability AI, which as a `cc-by-sa-4.0` license.
The fine-tuning data was mostly generated by OpenAI API calls to gpt-4, via [airoboros](https://github.com/jondurbin/airoboros)
The ToS for OpenAI API usage has a clause preventing the output from being used to train a model that __competes__ with OpenAI
- what does *compete* actually mean here?
- these small open source models will not produce output anywhere near the quality of gpt-4, or even gpt-3.5, so I can't imagine this could credibly be considered competing in the first place
- if someone else uses the dataset to do the same, they wouldn't necessarily be violating the ToS because they didn't call the API, so I don't know how that works
- the training data used in essentially all large language models includes a significant amount of copyrighted or otherwise non-permissive licensing in the first place
- other work using the self-instruct method, e.g. the original here: https://github.com/yizhongw/self-instruct released the data and model as apache-2
I am purposingly leaving this license ambiguous (other than the fact you must comply with the Meta original license for llama-2) because I am not a lawyer and refuse to attempt to interpret all of the terms accordingly.
Your best bet is probably to avoid using this commercially due to the OpenAI API usage.
Either way, by using this model, you agree to completely indemnify me.
| [
"QUESTION_ANSWERING",
"SUMMARIZATION"
] | Non_BioNLP |
chienweichang/formatted_address | chienweichang | text2text-generation | [
"transformers",
"tensorboard",
"safetensors",
"mt5",
"text2text-generation",
"generated_from_trainer",
"dataset:cwchang/tw_address_large",
"base_model:google/mt5-small",
"base_model:finetune:google/mt5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,702,956,972,000 | 2023-12-19T04:49:04 | 92 | 0 | ---
base_model: google/mt5-small
datasets:
- cwchang/tw_address_large
license: apache-2.0
metrics:
- rouge
tags:
- generated_from_trainer
model-index:
- name: formatted_address
results:
- task:
type: summarization
name: Summarization
dataset:
name: cwchang/tw_address_large
type: cwchang/tw_address_large
metrics:
- type: rouge
value: 97.0
name: Rouge1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# formatted_address
This model is a fine-tuned version of [google/mt5-small](https://huggingface.co/google/mt5-small) on the cwchang/tw_address_large dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1388
- Rouge1: 97.0
- Rouge2: 48.3471
- Rougel: 96.996
- Rougelsum: 96.9932
- Gen Len: 13.7152
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 6
- eval_batch_size: 6
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
### Framework versions
- Transformers 4.37.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
| [
"SUMMARIZATION"
] | Non_BioNLP |
am-azadi/gte-multilingual-base_Fine_Tuned_1e | am-azadi | sentence-similarity | [
"sentence-transformers",
"safetensors",
"new",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:25743",
"loss:MultipleNegativesRankingLoss",
"custom_code",
"arxiv:1908.10084",
"arxiv:1705.00652",
"base_model:Alibaba-NLP/gte-multilingual-base",
"base_model:finetune:Alibaba-NLP/gte-multilingual-base",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,740,074,285,000 | 2025-02-20T17:58:47 | 11 | 0 | ---
base_model: Alibaba-NLP/gte-multilingual-base
library_name: sentence-transformers
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:25743
- loss:MultipleNegativesRankingLoss
widget:
- source_sentence: م الحين SHIA WAVES ENGLISH Indians throw thousands of idols on
the street for not protecting them against the virus |Many people in India have
thrown away statues of gods and blamed why the gods with infinite power cannot
protect them from the ravages of the coronavirus? All sects must be repositioned,
otherwise there will be a bigger crisis in each sect. . . I don't know, when will
it be our country's turn? you say?
sentences:
- Esta mulher sofreu uma convulsão após ser vacinada contra a covid-19 na Argentina
- Images of Hindu idols destroyed for not protecting Indian people during the Covid-19
pandemic
- Forces raid a house in Indian-administered Kashmir
- source_sentence: 'En el mismo cuerpo legal atacaremos la raíz del problema: los
jefes de las mafias. Tipificaremos el nuevo delito de “autoría por dominio de
organización”. Es decir: los jefes de las bandas pagarán también por los delitos
que ordenen cometer a sus cómplices.'
sentences:
- Walmart va demander une preuve de vaccination à ses clients canadiens
- Vídeo mostra fraude de mortes na pandemia de Covid-19
- La autoría por dominio de organización sería un nuevo delito en Ecuador
- source_sentence: Winning
sentences:
- President Donald Trump has 232 electoral votes, Joe Biden has 212, 226 or 227.
- Suspected drunk drivers automatically face one month in jail under new law in
Thailand?
- Le bilan des violences post-électorales à M'Batto a atteint au moins une trentaine
de morts
- source_sentence: Pablo Iglesias Iglesias_ No soy partidario de la violencia pero
disfrutaría viendo como matan a tiros a los líderes del PP. La derecha debe ser
exterminada como un virus. 11:26 AM 24 ene. 12 1.682 Retweets 2.069 Likes 27 go
sentences:
- Pablo Iglesias tuiteó que disfrutaría de ver como matan de un tiro a líderes del
PP y a la derecha española habría que exterminarla como a un virus
- Delfines en un puerto de España durante el confinamiento
- Jenazah korban virus corona di Rusia
- source_sentence: 'ليس داعشياً من بيده المسدس ..انه جندي فرنسي ينفذ اعدامات بحق مواطنين
عزل في الجزائر !!! لم يكن حينها لا تنظيم قاعدة ولا دولة اسلامية ولا نصرة ليلصقوا
بهم منفردين تهمة الارهاب !! انتم ام واب واخ وابن وجد الارهاب .. Not Daashaa of
the pistol in his hand .. he''s a French soldier executions carried out against
unarmed civilians in Algeria !!! If not then it does not regulate not base an
Islamic state nor a victory for Alsqoa their individual terrorism charge !! You
are a mother and father and brother and the son of terror found .. Non Daashaa
du pistolet dans sa main .. Il est un soldat français exécutions menées contre
des civils non armés en Algérie !!! Si non, alors il ne réglemente pas pas fonder
un Etat islamique, ni une victoire pour Alsqoa leur charge individuelle du terrorisme
!! Vous êtes une mère et père et le frère et le fils de la terreur trouvé .. #
occupant'
sentences:
- Massacre perpétré par des soldats français en Algérie
- Video Of Attack On UP Minister Shrikant Sharma
- Map shows there are no wildfires in Canada and Mexico
---
# SentenceTransformer based on Alibaba-NLP/gte-multilingual-base
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Alibaba-NLP/gte-multilingual-base](https://huggingface.co/Alibaba-NLP/gte-multilingual-base). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [Alibaba-NLP/gte-multilingual-base](https://huggingface.co/Alibaba-NLP/gte-multilingual-base) <!-- at revision ca1791e0bcc104f6db161f27de1340241b13c5a4 -->
- **Maximum Sequence Length:** 8192 tokens
- **Output Dimensionality:** 768 dimensions
- **Similarity Function:** Cosine Similarity
<!-- - **Training Dataset:** Unknown -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 8192, 'do_lower_case': False}) with Transformer model: NewModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
"ليس داعشياً من بيده المسدس ..انه جندي فرنسي ينفذ اعدامات بحق مواطنين عزل في الجزائر !!! لم يكن حينها لا تنظيم قاعدة ولا دولة اسلامية ولا نصرة ليلصقوا بهم منفردين تهمة الارهاب !! انتم ام واب واخ وابن وجد الارهاب .. Not Daashaa of the pistol in his hand .. he's a French soldier executions carried out against unarmed civilians in Algeria !!! If not then it does not regulate not base an Islamic state nor a victory for Alsqoa their individual terrorism charge !! You are a mother and father and brother and the son of terror found .. Non Daashaa du pistolet dans sa main .. Il est un soldat français exécutions menées contre des civils non armés en Algérie !!! Si non, alors il ne réglemente pas pas fonder un Etat islamique, ni une victoire pour Alsqoa leur charge individuelle du terrorisme !! Vous êtes une mère et père et le frère et le fils de la terreur trouvé .. # occupant",
'Massacre perpétré par des soldats français en Algérie',
'Video Of Attack On UP Minister Shrikant Sharma',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### Unnamed Dataset
* Size: 25,743 training samples
* Columns: <code>sentence_0</code>, <code>sentence_1</code>, and <code>label</code>
* Approximate statistics based on the first 1000 samples:
| | sentence_0 | sentence_1 | label |
|:--------|:-------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:--------------------------------------------------------------|
| type | string | string | float |
| details | <ul><li>min: 2 tokens</li><li>mean: 140.38 tokens</li><li>max: 2514 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 20.49 tokens</li><li>max: 141 tokens</li></ul> | <ul><li>min: 1.0</li><li>mean: 1.0</li><li>max: 1.0</li></ul> |
* Samples:
| sentence_0 | sentence_1 | label |
|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------|:-----------------|
| <code>Olhem aí a mineradora da Noruega destruindo o meio ambiente na Amazônia. Lula vendeu o solo para a Noruega em documento secreto. Ela arrecada 2 bilhoes ao ano e devolve 180 milhoes para consertar o estrago que ela mesmo faz na Amazônia.</code> | <code>O ex-presidente Lula vendeu o solo da Amazônia para uma empresa norueguesa</code> | <code>1.0</code> |
| <code>EL CONGRESO DANIE Cometió una burrada Al aprobar en primera votación con 113 votos a favor, 5 en contra y una abstención, que la vacuna contra el coronavirus sea de manera OBLIGATORIA para todos Que les pasa a estos genios de la política, acaso no saben que están violando leyes universales de Derechos Humanos¿Qué les pasa a estos congresistas?. . ¿ Acaso desconocen y pisotean las leyes internacionales que respaldan los Derechos Humanos Universales ???. . Absolutamente nadie puede ser obligado a vacunarse. . Igualmente, ningún procedimiento médico puede hacerse sin el consentimiento del paciente. . No lo digo yo, lo dice la UNESCO,la Organización de las Naciones Unidas para la Educación, la Ciencia y la Cultura.... Que en sus normativas explican lo siguiente : . SOLO UNO MISMO TIENE EL CONTROL DE SU PROPIO CUERPO, nadie tiene el control de nuestro cuerpo más que uno mismo, nadie puede intervenir en nuestro cuerpo bajo ninguna circunstancia sin nuestro consentimiento. . Legalmente bajo t...</code> | <code>En Perú el Congreso aprobó que la vacuna contra el covid-19 sea obligatoria</code> | <code>1.0</code> |
| <code>Why changes to Legislation is so difficult. Debating PTSD in Emergency Services Debating Mental Health Stigma Debating Workers Compensation Debating Cancer Legislation for Firefighters Debating MP's Pay Debating PFAS Contamination Debating Suicide Figures in Australia Debating MP's AllowancesThis tells us everything we need to know about this Government’s priorities.</code> | <code>Accurate description of photos showing the difference in attendance in various parliamentary sessions in Australia</code> | <code>1.0</code> |
* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 20.0,
"similarity_fct": "cos_sim"
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `per_device_train_batch_size`: 1
- `per_device_eval_batch_size`: 1
- `num_train_epochs`: 1
- `multi_dataset_batch_sampler`: round_robin
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: no
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 1
- `per_device_eval_batch_size`: 1
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 1
- `eval_accumulation_steps`: None
- `torch_empty_cache_steps`: None
- `learning_rate`: 5e-05
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1
- `num_train_epochs`: 1
- `max_steps`: -1
- `lr_scheduler_type`: linear
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.0
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: False
- `fp16`: False
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: None
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: False
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: None
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `include_for_metrics`: []
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `dispatch_batches`: None
- `split_batches`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `eval_on_start`: False
- `use_liger_kernel`: False
- `eval_use_gather_object`: False
- `average_tokens_across_devices`: False
- `prompts`: None
- `batch_sampler`: batch_sampler
- `multi_dataset_batch_sampler`: round_robin
</details>
### Training Logs
| Epoch | Step | Training Loss |
|:------:|:-----:|:-------------:|
| 0.0194 | 500 | 0.0 |
| 0.0388 | 1000 | 0.0 |
| 0.0583 | 1500 | 0.0 |
| 0.0777 | 2000 | 0.0 |
| 0.0971 | 2500 | 0.0 |
| 0.1165 | 3000 | 0.0 |
| 0.1360 | 3500 | 0.0 |
| 0.1554 | 4000 | 0.0 |
| 0.1748 | 4500 | 0.0 |
| 0.1942 | 5000 | 0.0 |
| 0.2137 | 5500 | 0.0 |
| 0.2331 | 6000 | 0.0 |
| 0.2525 | 6500 | 0.0 |
| 0.2719 | 7000 | 0.0 |
| 0.2913 | 7500 | 0.0 |
| 0.3108 | 8000 | 0.0 |
| 0.3302 | 8500 | 0.0 |
| 0.3496 | 9000 | 0.0 |
| 0.3690 | 9500 | 0.0 |
| 0.3885 | 10000 | 0.0 |
| 0.4079 | 10500 | 0.0 |
| 0.4273 | 11000 | 0.0 |
| 0.4467 | 11500 | 0.0 |
| 0.4661 | 12000 | 0.0 |
| 0.4856 | 12500 | 0.0 |
| 0.5050 | 13000 | 0.0 |
| 0.5244 | 13500 | 0.0 |
| 0.5438 | 14000 | 0.0 |
| 0.5633 | 14500 | 0.0 |
| 0.5827 | 15000 | 0.0 |
| 0.6021 | 15500 | 0.0 |
| 0.6215 | 16000 | 0.0 |
| 0.6410 | 16500 | 0.0 |
| 0.6604 | 17000 | 0.0 |
| 0.6798 | 17500 | 0.0 |
| 0.6992 | 18000 | 0.0 |
| 0.7186 | 18500 | 0.0 |
| 0.7381 | 19000 | 0.0 |
| 0.7575 | 19500 | 0.0 |
| 0.7769 | 20000 | 0.0 |
| 0.7963 | 20500 | 0.0 |
| 0.8158 | 21000 | 0.0 |
| 0.8352 | 21500 | 0.0 |
| 0.8546 | 22000 | 0.0 |
| 0.8740 | 22500 | 0.0 |
| 0.8934 | 23000 | 0.0 |
| 0.9129 | 23500 | 0.0 |
| 0.9323 | 24000 | 0.0 |
| 0.9517 | 24500 | 0.0 |
| 0.9711 | 25000 | 0.0 |
| 0.9906 | 25500 | 0.0 |
### Framework Versions
- Python: 3.11.11
- Sentence Transformers: 3.4.1
- Transformers: 4.48.3
- PyTorch: 2.5.1+cu124
- Accelerate: 1.3.0
- Datasets: 3.3.1
- Tokenizers: 0.21.0
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### MultipleNegativesRankingLoss
```bibtex
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
rifatul123/Primary_doctor_v1 | rifatul123 | text-generation | [
"adapter-transformers",
"pytorch",
"gpt2",
"biology",
"medical",
"chemistry",
"text-generation-inference",
"text-generation",
"en",
"region:us"
] | 1,683,275,744,000 | 2023-05-05T16:57:39 | 0 | 0 | ---
language:
- en
library_name: adapter-transformers
metrics:
- accuracy
pipeline_tag: text-generation
tags:
- biology
- medical
- chemistry
- text-generation-inference
---
# Model Card for Model ID
This modelcard describes a fine-tuned GPT-2 language model for medical research using a personally collected dataset.
The model is intended for text generation in the medical research domain.
## Model Details
This modelcard describes a fine-tuned GPT-2 language model for medical research using a personally collected dataset. The model is intended for text generation in the medical research domain.
### Model Description
The model has been fine-tuned on a GPT-2 architecture and trained with a task-specific parameter for text generation. The do_sample parameter is set to true, which means that the model can generate text on its own rather than simply copying from the input. The max_length parameter is set to 50, which means that the maximum length of the generated text will be 50 tokens.
- **Developed by:** [OpenAI]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [Language Model]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [GPT-2]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
### Direct Use
This model can be used for text generation in the medical research domain. It can be used to generate text for a variety of purposes, such as research papers, reports, and summaries.
### Downstream Use [optional]
The model can be fine-tuned for downstream tasks such as summarization, question answering, and text classification.
### Out-of-Scope Use
This model may not perform as well on text outside the medical research domain. It is important to carefully evaluate the generated text to ensure that it is appropriate for the intended use.
## Bias, Risks, and Limitations
This modelcard acknowledges that all language models have limitations and potential biases. The model may produce biased or inaccurate outputs if the input data contains bias or if the training data is not diverse enough. The risks of using the model include the possibility of generating misleading or harmful information.
### Recommendations
To mitigate potential risks and limitations, users of the model should carefully evaluate the generated text and consider the following recommendations:
1)Evaluate the input data for potential bias and ensure that it is diverse and representative.
2)Consider fine-tuning the model on additional data to improve its accuracy and reduce the risk of bias.
3)Review and edit the generated text before use to ensure that it is appropriate for the intended purpose.
4)Provide clear and transparent documentation of the model's limitations and potential biases to users and stakeholders.
## How to Get Started with the Model
To use the model, load it in your preferred programming language using the transformers library, and pass in the input text. The model will generate text based on the input, using the task-specific parameters that have been set.
| [
"TEXT_CLASSIFICATION",
"QUESTION_ANSWERING",
"SUMMARIZATION"
] | BioNLP |
Helsinki-NLP/opus-mt-yo-fr | Helsinki-NLP | translation | [
"transformers",
"pytorch",
"tf",
"marian",
"text2text-generation",
"translation",
"yo",
"fr",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,646,263,744,000 | 2023-08-16T12:09:04 | 57 | 0 | ---
license: apache-2.0
tags:
- translation
---
### opus-mt-yo-fr
* source languages: yo
* target languages: fr
* OPUS readme: [yo-fr](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/yo-fr/README.md)
* dataset: opus
* model: transformer-align
* pre-processing: normalization + SentencePiece
* download original weights: [opus-2020-01-16.zip](https://object.pouta.csc.fi/OPUS-MT-models/yo-fr/opus-2020-01-16.zip)
* test set translations: [opus-2020-01-16.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/yo-fr/opus-2020-01-16.test.txt)
* test set scores: [opus-2020-01-16.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/yo-fr/opus-2020-01-16.eval.txt)
## Benchmarks
| testset | BLEU | chr-F |
|-----------------------|-------|-------|
| JW300.yo.fr | 24.1 | 0.408 |
| [
"TRANSLATION"
] | Non_BioNLP |
gokuls/mobilebert_sa_GLUE_Experiment_logit_kd_stsb_256 | gokuls | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"mobilebert",
"text-classification",
"generated_from_trainer",
"en",
"dataset:glue",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,675,061,383,000 | 2023-01-30T06:53:58 | 138 | 0 | ---
datasets:
- glue
language:
- en
license: apache-2.0
metrics:
- spearmanr
tags:
- generated_from_trainer
model-index:
- name: mobilebert_sa_GLUE_Experiment_logit_kd_stsb_256
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: GLUE STSB
type: glue
config: stsb
split: validation
args: stsb
metrics:
- type: spearmanr
value: 0.016625925233910453
name: Spearmanr
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# mobilebert_sa_GLUE_Experiment_logit_kd_stsb_256
This model is a fine-tuned version of [google/mobilebert-uncased](https://huggingface.co/google/mobilebert-uncased) on the GLUE STSB dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1337
- Pearson: 0.0151
- Spearmanr: 0.0166
- Combined Score: 0.0159
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 128
- eval_batch_size: 128
- seed: 10
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss | Pearson | Spearmanr | Combined Score |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:---------:|:--------------:|
| 2.075 | 1.0 | 45 | 1.1337 | 0.0151 | 0.0166 | 0.0159 |
| 1.0752 | 2.0 | 90 | 1.1691 | 0.0603 | 0.0648 | 0.0626 |
| 1.0435 | 3.0 | 135 | 1.2035 | 0.0659 | 0.0746 | 0.0703 |
| 1.0472 | 4.0 | 180 | 1.1488 | 0.0764 | 0.0817 | 0.0790 |
| 0.9687 | 5.0 | 225 | 1.5234 | 0.0979 | 0.0959 | 0.0969 |
| 0.9016 | 6.0 | 270 | 1.2243 | 0.1434 | 0.1381 | 0.1408 |
### Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
Helsinki-NLP/opus-mt-tc-bible-big-itc-deu_eng_fra_por_spa | Helsinki-NLP | translation | [
"transformers",
"pytorch",
"safetensors",
"marian",
"text2text-generation",
"translation",
"opus-mt-tc-bible",
"acf",
"an",
"ast",
"ca",
"cbk",
"co",
"crs",
"de",
"egl",
"en",
"es",
"ext",
"fr",
"frm",
"fro",
"frp",
"fur",
"gcf",
"gl",
"ht",
"it",
"kea",
"la",
"lad",
"lij",
"lld",
"lmo",
"lou",
"mfe",
"mo",
"mwl",
"nap",
"oc",
"osp",
"pap",
"pcd",
"pms",
"pt",
"rm",
"ro",
"rup",
"sc",
"scn",
"vec",
"wa",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,728,377,856,000 | 2024-10-08T08:57:47 | 116 | 0 | ---
language:
- acf
- an
- ast
- ca
- cbk
- co
- crs
- de
- egl
- en
- es
- ext
- fr
- frm
- fro
- frp
- fur
- gcf
- gl
- ht
- it
- kea
- la
- lad
- lij
- lld
- lmo
- lou
- mfe
- mo
- mwl
- nap
- oc
- osp
- pap
- pcd
- pms
- pt
- rm
- ro
- rup
- sc
- scn
- vec
- wa
library_name: transformers
license: apache-2.0
tags:
- translation
- opus-mt-tc-bible
model-index:
- name: opus-mt-tc-bible-big-itc-deu_eng_fra_por_spa
results:
- task:
type: translation
name: Translation ast-deu
dataset:
name: flores200-devtest
type: flores200-devtest
args: ast-deu
metrics:
- type: bleu
value: 24.8
name: BLEU
- type: chrf
value: 0.53776
name: chr-F
- type: bleu
value: 36.8
name: BLEU
- type: chrf
value: 0.61482
name: chr-F
- type: bleu
value: 31.3
name: BLEU
- type: chrf
value: 0.56504
name: chr-F
- type: bleu
value: 31.1
name: BLEU
- type: chrf
value: 0.57158
name: chr-F
- type: bleu
value: 21.2
name: BLEU
- type: chrf
value: 0.49579
name: chr-F
- type: bleu
value: 29.2
name: BLEU
- type: chrf
value: 0.58203
name: chr-F
- type: bleu
value: 44.6
name: BLEU
- type: chrf
value: 0.69165
name: chr-F
- type: bleu
value: 38.9
name: BLEU
- type: chrf
value: 0.63612
name: chr-F
- type: bleu
value: 37.7
name: BLEU
- type: chrf
value: 0.62911
name: chr-F
- type: bleu
value: 24.6
name: BLEU
- type: chrf
value: 0.5332
name: chr-F
- type: bleu
value: 29.1
name: BLEU
- type: chrf
value: 0.58592
name: chr-F
- type: bleu
value: 43.8
name: BLEU
- type: chrf
value: 0.68067
name: chr-F
- type: bleu
value: 37.0
name: BLEU
- type: chrf
value: 0.62388
name: chr-F
- type: bleu
value: 24.4
name: BLEU
- type: chrf
value: 0.52983
name: chr-F
- type: bleu
value: 21.8
name: BLEU
- type: chrf
value: 0.51969
name: chr-F
- type: bleu
value: 34.3
name: BLEU
- type: chrf
value: 0.60793
name: chr-F
- type: bleu
value: 30.0
name: BLEU
- type: chrf
value: 0.56989
name: chr-F
- type: bleu
value: 29.3
name: BLEU
- type: chrf
value: 0.56207
name: chr-F
- type: bleu
value: 20.0
name: BLEU
- type: chrf
value: 0.48436
name: chr-F
- type: bleu
value: 27.7
name: BLEU
- type: chrf
value: 0.57369
name: chr-F
- type: bleu
value: 40.0
name: BLEU
- type: chrf
value: 0.66358
name: chr-F
- type: bleu
value: 36.5
name: BLEU
- type: chrf
value: 0.62487
name: chr-F
- type: bleu
value: 32.7
name: BLEU
- type: chrf
value: 0.60267
name: chr-F
- type: bleu
value: 24.3
name: BLEU
- type: chrf
value: 0.53227
name: chr-F
- type: bleu
value: 19.1
name: BLEU
- type: chrf
value: 0.49916
name: chr-F
- type: bleu
value: 32.5
name: BLEU
- type: chrf
value: 0.59656
name: chr-F
- type: bleu
value: 35.4
name: BLEU
- type: chrf
value: 0.61574
name: chr-F
- type: bleu
value: 27.7
name: BLEU
- type: chrf
value: 0.55195
name: chr-F
- type: bleu
value: 18.4
name: BLEU
- type: chrf
value: 0.47382
name: chr-F
- type: bleu
value: 24.1
name: BLEU
- type: chrf
value: 0.55779
name: chr-F
- type: bleu
value: 32.2
name: BLEU
- type: chrf
value: 0.61563
name: chr-F
- type: bleu
value: 31.2
name: BLEU
- type: chrf
value: 0.6021
name: chr-F
- type: bleu
value: 28.8
name: BLEU
- type: chrf
value: 0.58279
name: chr-F
- type: bleu
value: 23.2
name: BLEU
- type: chrf
value: 0.52348
name: chr-F
- type: bleu
value: 19.3
name: BLEU
- type: chrf
value: 0.49089
name: chr-F
- type: bleu
value: 35.5
name: BLEU
- type: chrf
value: 0.60553
name: chr-F
- type: bleu
value: 26.6
name: BLEU
- type: chrf
value: 0.54027
name: chr-F
- type: bleu
value: 28.9
name: BLEU
- type: chrf
value: 0.57696
name: chr-F
- type: bleu
value: 18.0
name: BLEU
- type: chrf
value: 0.46974
name: chr-F
- type: bleu
value: 22.7
name: BLEU
- type: chrf
value: 0.51695
name: chr-F
- type: bleu
value: 36.2
name: BLEU
- type: chrf
value: 0.62347
name: chr-F
- type: bleu
value: 31.4
name: BLEU
- type: chrf
value: 0.57498
name: chr-F
- type: bleu
value: 29.4
name: BLEU
- type: chrf
value: 0.56183
name: chr-F
- type: bleu
value: 20.0
name: BLEU
- type: chrf
value: 0.48038
name: chr-F
- type: bleu
value: 15.4
name: BLEU
- type: chrf
value: 0.45516
name: chr-F
- type: bleu
value: 25.5
name: BLEU
- type: chrf
value: 0.5354
name: chr-F
- type: bleu
value: 22.2
name: BLEU
- type: chrf
value: 0.50076
name: chr-F
- type: bleu
value: 22.9
name: BLEU
- type: chrf
value: 0.50134
name: chr-F
- type: bleu
value: 16.2
name: BLEU
- type: chrf
value: 0.44053
name: chr-F
- type: bleu
value: 28.7
name: BLEU
- type: chrf
value: 0.57822
name: chr-F
- type: bleu
value: 50.7
name: BLEU
- type: chrf
value: 0.7303
name: chr-F
- type: bleu
value: 39.7
name: BLEU
- type: chrf
value: 0.649
name: chr-F
- type: bleu
value: 36.9
name: BLEU
- type: chrf
value: 0.63318
name: chr-F
- type: bleu
value: 22.9
name: BLEU
- type: chrf
value: 0.52269
name: chr-F
- type: bleu
value: 23.2
name: BLEU
- type: chrf
value: 0.53166
name: chr-F
- type: bleu
value: 44.6
name: BLEU
- type: chrf
value: 0.68541
name: chr-F
- type: bleu
value: 30.5
name: BLEU
- type: chrf
value: 0.57224
name: chr-F
- type: bleu
value: 33.2
name: BLEU
- type: chrf
value: 0.59064
name: chr-F
- type: bleu
value: 21.7
name: BLEU
- type: chrf
value: 0.49601
name: chr-F
- type: bleu
value: 30.3
name: BLEU
- type: chrf
value: 0.59047
name: chr-F
- type: bleu
value: 48.0
name: BLEU
- type: chrf
value: 0.71096
name: chr-F
- type: bleu
value: 40.1
name: BLEU
- type: chrf
value: 0.64555
name: chr-F
- type: bleu
value: 25.1
name: BLEU
- type: chrf
value: 0.534
name: chr-F
- type: bleu
value: 28.7
name: BLEU
- type: chrf
value: 0.58428
name: chr-F
- type: bleu
value: 41.8
name: BLEU
- type: chrf
value: 0.67719
name: chr-F
- type: bleu
value: 37.6
name: BLEU
- type: chrf
value: 0.63678
name: chr-F
- type: bleu
value: 36.1
name: BLEU
- type: chrf
value: 0.62371
name: chr-F
- type: bleu
value: 24.5
name: BLEU
- type: chrf
value: 0.5315
name: chr-F
- type: bleu
value: 19.2
name: BLEU
- type: chrf
value: 0.48102
name: chr-F
- type: bleu
value: 29.6
name: BLEU
- type: chrf
value: 0.55782
name: chr-F
- type: bleu
value: 26.1
name: BLEU
- type: chrf
value: 0.52773
name: chr-F
- type: bleu
value: 25.2
name: BLEU
- type: chrf
value: 0.51894
name: chr-F
- type: bleu
value: 17.9
name: BLEU
- type: chrf
value: 0.45724
name: chr-F
- type: bleu
value: 21.5
name: BLEU
- type: chrf
value: 0.53451
name: chr-F
- type: bleu
value: 28.5
name: BLEU
- type: chrf
value: 0.58896
name: chr-F
- type: bleu
value: 27.6
name: BLEU
- type: chrf
value: 0.57406
name: chr-F
- type: bleu
value: 25.2
name: BLEU
- type: chrf
value: 0.55749
name: chr-F
- type: bleu
value: 19.9
name: BLEU
- type: chrf
value: 0.49238
name: chr-F
- type: bleu
value: 34.2
name: BLEU
- type: chrf
value: 0.59392
name: chr-F
- type: bleu
value: 27.6
name: BLEU
- type: chrf
value: 0.54003
name: chr-F
- type: bleu
value: 27.9
name: BLEU
- type: chrf
value: 0.53842
name: chr-F
- type: bleu
value: 18.2
name: BLEU
- type: chrf
value: 0.46002
name: chr-F
- type: bleu
value: 19.3
name: BLEU
- type: chrf
value: 0.48795
name: chr-F
- type: bleu
value: 30.7
name: BLEU
- type: chrf
value: 0.5684
name: chr-F
- type: bleu
value: 27.3
name: BLEU
- type: chrf
value: 0.54164
name: chr-F
- type: bleu
value: 26.2
name: BLEU
- type: chrf
value: 0.53482
name: chr-F
- type: bleu
value: 18.4
name: BLEU
- type: chrf
value: 0.46588
name: chr-F
- task:
type: translation
name: Translation ast-deu
dataset:
name: flores101-devtest
type: flores_101
args: ast deu devtest
metrics:
- type: bleu
value: 24.2
name: BLEU
- type: chrf
value: 0.53243
name: chr-F
- type: bleu
value: 36.0
name: BLEU
- type: chrf
value: 0.61235
name: chr-F
- type: bleu
value: 31.2
name: BLEU
- type: chrf
value: 0.56687
name: chr-F
- type: bleu
value: 30.6
name: BLEU
- type: chrf
value: 0.57033
name: chr-F
- type: bleu
value: 21.2
name: BLEU
- type: chrf
value: 0.49637
name: chr-F
- type: bleu
value: 38.4
name: BLEU
- type: chrf
value: 0.63271
name: chr-F
- type: bleu
value: 28.9
name: BLEU
- type: chrf
value: 0.58433
name: chr-F
- type: bleu
value: 43.3
name: BLEU
- type: chrf
value: 0.67826
name: chr-F
- type: bleu
value: 27.1
name: BLEU
- type: chrf
value: 0.56897
name: chr-F
- type: bleu
value: 24.2
name: BLEU
- type: chrf
value: 0.53183
name: chr-F
- type: bleu
value: 28.4
name: BLEU
- type: chrf
value: 0.57961
name: chr-F
- type: bleu
value: 18.3
name: BLEU
- type: chrf
value: 0.48105
name: chr-F
- type: bleu
value: 35.0
name: BLEU
- type: chrf
value: 0.60362
name: chr-F
- type: bleu
value: 29.0
name: BLEU
- type: chrf
value: 0.57808
name: chr-F
- type: bleu
value: 17.6
name: BLEU
- type: chrf
value: 0.46648
name: chr-F
- type: bleu
value: 28.0
name: BLEU
- type: chrf
value: 0.57391
name: chr-F
- type: bleu
value: 49.4
name: BLEU
- type: chrf
value: 0.72351
name: chr-F
- type: bleu
value: 47.4
name: BLEU
- type: chrf
value: 0.70724
name: chr-F
- type: bleu
value: 39.2
name: BLEU
- type: chrf
value: 0.64103
name: chr-F
- type: bleu
value: 25.0
name: BLEU
- type: chrf
value: 0.53268
name: chr-F
- type: bleu
value: 28.1
name: BLEU
- type: chrf
value: 0.5798
name: chr-F
- type: bleu
value: 41.6
name: BLEU
- type: chrf
value: 0.67583
name: chr-F
- type: bleu
value: 24.3
name: BLEU
- type: chrf
value: 0.53082
name: chr-F
- type: bleu
value: 27.1
name: BLEU
- type: chrf
value: 0.57039
name: chr-F
- type: bleu
value: 25.0
name: BLEU
- type: chrf
value: 0.55607
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: generaltest2022
type: generaltest2022
args: fra-deu
metrics:
- type: bleu
value: 42.4
name: BLEU
- type: chrf
value: 0.66476
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: multi30k_test_2016_flickr
type: multi30k-2016_flickr
args: fra-deu
metrics:
- type: bleu
value: 32.6
name: BLEU
- type: chrf
value: 0.61797
name: chr-F
- type: bleu
value: 47.2
name: BLEU
- type: chrf
value: 0.66271
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: multi30k_test_2017_flickr
type: multi30k-2017_flickr
args: fra-deu
metrics:
- type: bleu
value: 29.4
name: BLEU
- type: chrf
value: 0.59701
name: chr-F
- type: bleu
value: 50.3
name: BLEU
- type: chrf
value: 0.69422
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: multi30k_test_2017_mscoco
type: multi30k-2017_mscoco
args: fra-deu
metrics:
- type: bleu
value: 25.7
name: BLEU
- type: chrf
value: 0.55509
name: chr-F
- type: bleu
value: 48.7
name: BLEU
- type: chrf
value: 0.67791
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: multi30k_test_2018_flickr
type: multi30k-2018_flickr
args: fra-deu
metrics:
- type: bleu
value: 24.0
name: BLEU
- type: chrf
value: 0.55237
name: chr-F
- type: bleu
value: 43.8
name: BLEU
- type: chrf
value: 0.64722
name: chr-F
- task:
type: translation
name: Translation fra-eng
dataset:
name: newsdiscusstest2015
type: newsdiscusstest2015
args: fra-eng
metrics:
- type: bleu
value: 38.4
name: BLEU
- type: chrf
value: 0.61385
name: chr-F
- task:
type: translation
name: Translation cat-deu
dataset:
name: ntrex128
type: ntrex128
args: cat-deu
metrics:
- type: bleu
value: 24.0
name: BLEU
- type: chrf
value: 0.54096
name: chr-F
- type: bleu
value: 36.5
name: BLEU
- type: chrf
value: 0.63516
name: chr-F
- type: bleu
value: 28.1
name: BLEU
- type: chrf
value: 0.56385
name: chr-F
- type: bleu
value: 28.7
name: BLEU
- type: chrf
value: 0.56246
name: chr-F
- type: bleu
value: 35.8
name: BLEU
- type: chrf
value: 0.61311
name: chr-F
- type: bleu
value: 23.4
name: BLEU
- type: chrf
value: 0.53059
name: chr-F
- type: bleu
value: 34.7
name: BLEU
- type: chrf
value: 0.61285
name: chr-F
- type: bleu
value: 25.8
name: BLEU
- type: chrf
value: 0.54075
name: chr-F
- type: bleu
value: 30.6
name: BLEU
- type: chrf
value: 0.56863
name: chr-F
- type: bleu
value: 23.6
name: BLEU
- type: chrf
value: 0.53724
name: chr-F
- type: bleu
value: 38.7
name: BLEU
- type: chrf
value: 0.64481
name: chr-F
- type: bleu
value: 27.8
name: BLEU
- type: chrf
value: 0.55856
name: chr-F
- type: bleu
value: 28.7
name: BLEU
- type: chrf
value: 0.56322
name: chr-F
- type: bleu
value: 36.8
name: BLEU
- type: chrf
value: 0.61794
name: chr-F
- type: bleu
value: 25.0
name: BLEU
- type: chrf
value: 0.54678
name: chr-F
- type: bleu
value: 39.2
name: BLEU
- type: chrf
value: 0.64636
name: chr-F
- type: bleu
value: 30.0
name: BLEU
- type: chrf
value: 0.57428
name: chr-F
- type: bleu
value: 29.7
name: BLEU
- type: chrf
value: 0.56858
name: chr-F
- type: bleu
value: 33.0
name: BLEU
- type: chrf
value: 0.58886
name: chr-F
- type: bleu
value: 24.6
name: BLEU
- type: chrf
value: 0.54833
name: chr-F
- type: bleu
value: 39.7
name: BLEU
- type: chrf
value: 0.65223
name: chr-F
- type: bleu
value: 28.9
name: BLEU
- type: chrf
value: 0.56793
name: chr-F
- type: bleu
value: 33.8
name: BLEU
- type: chrf
value: 0.59218
name: chr-F
- type: bleu
value: 22.4
name: BLEU
- type: chrf
value: 0.53249
name: chr-F
- type: bleu
value: 33.8
name: BLEU
- type: chrf
value: 0.61807
name: chr-F
- type: bleu
value: 26.4
name: BLEU
- type: chrf
value: 0.55575
name: chr-F
- type: bleu
value: 27.2
name: BLEU
- type: chrf
value: 0.55086
name: chr-F
- type: bleu
value: 31.9
name: BLEU
- type: chrf
value: 0.57787
name: chr-F
- type: bleu
value: 23.8
name: BLEU
- type: chrf
value: 0.54309
name: chr-F
- type: bleu
value: 37.4
name: BLEU
- type: chrf
value: 0.64416
name: chr-F
- type: bleu
value: 29.4
name: BLEU
- type: chrf
value: 0.5732
name: chr-F
- type: bleu
value: 29.0
name: BLEU
- type: chrf
value: 0.56751
name: chr-F
- task:
type: translation
name: Translation cat-deu
dataset:
name: tatoeba-test-v2021-08-07
type: tatoeba_mt
args: cat-deu
metrics:
- type: bleu
value: 47.9
name: BLEU
- type: chrf
value: 0.66856
name: chr-F
- type: bleu
value: 57.9
name: BLEU
- type: chrf
value: 0.72313
name: chr-F
- type: bleu
value: 53.8
name: BLEU
- type: chrf
value: 0.71565
name: chr-F
- type: bleu
value: 58.7
name: BLEU
- type: chrf
value: 0.75797
name: chr-F
- type: bleu
value: 77.7
name: BLEU
- type: chrf
value: 0.8761
name: chr-F
- type: bleu
value: 50.0
name: BLEU
- type: chrf
value: 0.68638
name: chr-F
- type: bleu
value: 58.0
name: BLEU
- type: chrf
value: 0.72664
name: chr-F
- type: bleu
value: 40.6
name: BLEU
- type: chrf
value: 0.62093
name: chr-F
- type: bleu
value: 52.0
name: BLEU
- type: chrf
value: 0.70764
name: chr-F
- type: bleu
value: 55.0
name: BLEU
- type: chrf
value: 0.72229
name: chr-F
- type: bleu
value: 55.7
name: BLEU
- type: chrf
value: 0.70552
name: chr-F
- type: bleu
value: 62.1
name: BLEU
- type: chrf
value: 0.77067
name: chr-F
- type: bleu
value: 72.1
name: BLEU
- type: chrf
value: 0.82795
name: chr-F
- type: bleu
value: 49.4
name: BLEU
- type: chrf
value: 0.68325
name: chr-F
- type: bleu
value: 70.5
name: BLEU
- type: chrf
value: 0.81176
name: chr-F
- type: bleu
value: 64.4
name: BLEU
- type: chrf
value: 0.78299
name: chr-F
- type: bleu
value: 55.6
name: BLEU
- type: chrf
value: 0.74169
name: chr-F
- type: bleu
value: 63.0
name: BLEU
- type: chrf
value: 0.77673
name: chr-F
- type: bleu
value: 36.7
name: BLEU
- type: chrf
value: 0.54247
name: chr-F
- type: bleu
value: 40.4
name: BLEU
- type: chrf
value: 0.5979
name: chr-F
- type: bleu
value: 24.8
name: BLEU
- type: chrf
value: 0.42548
name: chr-F
- type: bleu
value: 24.3
name: BLEU
- type: chrf
value: 0.42385
name: chr-F
- type: bleu
value: 25.2
name: BLEU
- type: chrf
value: 0.45821
name: chr-F
- type: bleu
value: 52.8
name: BLEU
- type: chrf
value: 0.69536
name: chr-F
- type: bleu
value: 22.4
name: BLEU
- type: chrf
value: 0.40921
name: chr-F
- type: bleu
value: 28.4
name: BLEU
- type: chrf
value: 0.49044
name: chr-F
- type: bleu
value: 20.8
name: BLEU
- type: chrf
value: 0.39308
name: chr-F
- type: bleu
value: 48.8
name: BLEU
- type: chrf
value: 0.68379
name: chr-F
- type: bleu
value: 64.2
name: BLEU
- type: chrf
value: 0.77089
name: chr-F
- type: bleu
value: 58.7
name: BLEU
- type: chrf
value: 0.75364
name: chr-F
- type: bleu
value: 50.3
name: BLEU
- type: chrf
value: 0.71396
name: chr-F
- type: bleu
value: 65.2
name: BLEU
- type: chrf
value: 0.79684
name: chr-F
- type: bleu
value: 50.3
name: BLEU
- type: chrf
value: 0.68217
name: chr-F
- type: bleu
value: 59.0
name: BLEU
- type: chrf
value: 0.73059
name: chr-F
- type: bleu
value: 54.1
name: BLEU
- type: chrf
value: 0.70724
name: chr-F
- type: bleu
value: 53.3
name: BLEU
- type: chrf
value: 0.73085
name: chr-F
- type: bleu
value: 57.6
name: BLEU
- type: chrf
value: 0.73813
name: chr-F
- type: bleu
value: 49.3
name: BLEU
- type: chrf
value: 0.68124
name: chr-F
- type: bleu
value: 61.0
name: BLEU
- type: chrf
value: 0.74977
name: chr-F
- type: bleu
value: 56.6
name: BLEU
- type: chrf
value: 0.73392
name: chr-F
- type: bleu
value: 61.1
name: BLEU
- type: chrf
value: 0.7728
name: chr-F
- type: bleu
value: 50.9
name: BLEU
- type: chrf
value: 0.68111
name: chr-F
- task:
type: translation
name: Translation fra-eng
dataset:
name: tico19-test
type: tico19-test
args: fra-eng
metrics:
- type: bleu
value: 39.7
name: BLEU
- type: chrf
value: 0.62364
name: chr-F
- type: bleu
value: 34.2
name: BLEU
- type: chrf
value: 0.58563
name: chr-F
- type: bleu
value: 36.5
name: BLEU
- type: chrf
value: 0.59556
name: chr-F
- type: bleu
value: 51.8
name: BLEU
- type: chrf
value: 0.7442
name: chr-F
- type: bleu
value: 34.5
name: BLEU
- type: chrf
value: 0.60081
name: chr-F
- type: bleu
value: 44.8
name: BLEU
- type: chrf
value: 0.68156
name: chr-F
- type: bleu
value: 50.3
name: BLEU
- type: chrf
value: 0.73454
name: chr-F
- type: bleu
value: 34.9
name: BLEU
- type: chrf
value: 0.60441
name: chr-F
- type: bleu
value: 42.7
name: BLEU
- type: chrf
value: 0.67749
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2008
type: wmt-2008-news
args: fra-deu
metrics:
- type: bleu
value: 22.9
name: BLEU
- type: chrf
value: 0.5318
name: chr-F
- type: bleu
value: 26.5
name: BLEU
- type: chrf
value: 0.54379
name: chr-F
- type: bleu
value: 33.1
name: BLEU
- type: chrf
value: 0.58804
name: chr-F
- type: bleu
value: 21.6
name: BLEU
- type: chrf
value: 0.52221
name: chr-F
- type: bleu
value: 27.9
name: BLEU
- type: chrf
value: 0.55331
name: chr-F
- type: bleu
value: 32.0
name: BLEU
- type: chrf
value: 0.58769
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2009
type: wmt-2009-news
args: fra-deu
metrics:
- type: bleu
value: 22.5
name: BLEU
- type: chrf
value: 0.52771
name: chr-F
- type: bleu
value: 30.2
name: BLEU
- type: chrf
value: 0.56679
name: chr-F
- type: bleu
value: 32.1
name: BLEU
- type: chrf
value: 0.58921
name: chr-F
- type: bleu
value: 22.8
name: BLEU
- type: chrf
value: 0.53022
name: chr-F
- type: bleu
value: 33.8
name: BLEU
- type: chrf
value: 0.59309
name: chr-F
- type: bleu
value: 32.0
name: BLEU
- type: chrf
value: 0.59309
name: chr-F
- type: bleu
value: 33.5
name: BLEU
- type: chrf
value: 0.5976
name: chr-F
- type: bleu
value: 22.3
name: BLEU
- type: chrf
value: 0.52822
name: chr-F
- type: bleu
value: 30.4
name: BLEU
- type: chrf
value: 0.56989
name: chr-F
- type: bleu
value: 32.2
name: BLEU
- type: chrf
value: 0.5915
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2010
type: wmt-2010-news
args: fra-deu
metrics:
- type: bleu
value: 24.0
name: BLEU
- type: chrf
value: 0.53765
name: chr-F
- type: bleu
value: 32.6
name: BLEU
- type: chrf
value: 0.59251
name: chr-F
- type: bleu
value: 37.6
name: BLEU
- type: chrf
value: 0.6248
name: chr-F
- type: bleu
value: 26.0
name: BLEU
- type: chrf
value: 0.55161
name: chr-F
- type: bleu
value: 36.3
name: BLEU
- type: chrf
value: 0.61562
name: chr-F
- type: bleu
value: 35.7
name: BLEU
- type: chrf
value: 0.62021
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2011
type: wmt-2011-news
args: fra-deu
metrics:
- type: bleu
value: 23.1
name: BLEU
- type: chrf
value: 0.53025
name: chr-F
- type: bleu
value: 32.9
name: BLEU
- type: chrf
value: 0.59636
name: chr-F
- type: bleu
value: 39.9
name: BLEU
- type: chrf
value: 0.63203
name: chr-F
- type: bleu
value: 23.3
name: BLEU
- type: chrf
value: 0.52934
name: chr-F
- type: bleu
value: 33.8
name: BLEU
- type: chrf
value: 0.59606
name: chr-F
- type: bleu
value: 34.9
name: BLEU
- type: chrf
value: 0.61079
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2012
type: wmt-2012-news
args: fra-deu
metrics:
- type: bleu
value: 24.0
name: BLEU
- type: chrf
value: 0.52957
name: chr-F
- type: bleu
value: 33.6
name: BLEU
- type: chrf
value: 0.59352
name: chr-F
- type: bleu
value: 39.2
name: BLEU
- type: chrf
value: 0.62641
name: chr-F
- type: bleu
value: 24.6
name: BLEU
- type: chrf
value: 0.53519
name: chr-F
- type: bleu
value: 37.4
name: BLEU
- type: chrf
value: 0.62284
name: chr-F
- type: bleu
value: 33.8
name: BLEU
- type: chrf
value: 0.61076
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2013
type: wmt-2013-news
args: fra-deu
metrics:
- type: bleu
value: 25.4
name: BLEU
- type: chrf
value: 0.54167
name: chr-F
- type: bleu
value: 34.0
name: BLEU
- type: chrf
value: 0.59236
name: chr-F
- type: bleu
value: 34.9
name: BLEU
- type: chrf
value: 0.59347
name: chr-F
- type: bleu
value: 26.3
name: BLEU
- type: chrf
value: 0.5513
name: chr-F
- type: bleu
value: 34.6
name: BLEU
- type: chrf
value: 0.60681
name: chr-F
- type: bleu
value: 33.2
name: BLEU
- type: chrf
value: 0.59816
name: chr-F
- task:
type: translation
name: Translation fra-eng
dataset:
name: newstest2014
type: wmt-2014-news
args: fra-eng
metrics:
- type: bleu
value: 37.9
name: BLEU
- type: chrf
value: 0.63499
name: chr-F
- task:
type: translation
name: Translation ron-eng
dataset:
name: newstest2016
type: wmt-2016-news
args: ron-eng
metrics:
- type: bleu
value: 39.5
name: BLEU
- type: chrf
value: 0.63996
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2019
type: wmt-2019-news
args: fra-deu
metrics:
- type: bleu
value: 28.6
name: BLEU
- type: chrf
value: 0.60468
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2020
type: wmt-2020-news
args: fra-deu
metrics:
- type: bleu
value: 28.8
name: BLEU
- type: chrf
value: 0.61401
name: chr-F
- task:
type: translation
name: Translation fra-deu
dataset:
name: newstest2021
type: wmt-2021-news
args: fra-deu
metrics:
- type: bleu
value: 39.5
name: BLEU
- type: chrf
value: 0.6595
name: chr-F
---
# opus-mt-tc-bible-big-itc-deu_eng_fra_por_spa
## Table of Contents
- [Model Details](#model-details)
- [Uses](#uses)
- [Risks, Limitations and Biases](#risks-limitations-and-biases)
- [How to Get Started With the Model](#how-to-get-started-with-the-model)
- [Training](#training)
- [Evaluation](#evaluation)
- [Citation Information](#citation-information)
- [Acknowledgements](#acknowledgements)
## Model Details
Neural machine translation model for translating from Italic languages (itc) to unknown (deu+eng+fra+por+spa).
This model is part of the [OPUS-MT project](https://github.com/Helsinki-NLP/Opus-MT), an effort to make neural machine translation models widely available and accessible for many languages in the world. All models are originally trained using the amazing framework of [Marian NMT](https://marian-nmt.github.io/), an efficient NMT implementation written in pure C++. The models have been converted to pyTorch using the transformers library by huggingface. Training data is taken from [OPUS](https://opus.nlpl.eu/) and training pipelines use the procedures of [OPUS-MT-train](https://github.com/Helsinki-NLP/Opus-MT-train).
**Model Description:**
- **Developed by:** Language Technology Research Group at the University of Helsinki
- **Model Type:** Translation (transformer-big)
- **Release**: 2024-05-30
- **License:** Apache-2.0
- **Language(s):**
- Source Language(s): acf arg ast cat cbk cos crs egl ext fra frm fro frp fur gcf glg hat ita kea lad lat lij lld lmo lou mfe mol mwl nap oci osp pap pcd pms por roh ron rup scn spa srd vec wln
- Target Language(s): deu eng fra por spa
- Valid Target Language Labels: >>deu<< >>eng<< >>fra<< >>por<< >>spa<< >>xxx<<
- **Original Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/itc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
- **Resources for more information:**
- [OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/itc-deu%2Beng%2Bfra%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
- [OPUS-MT-train GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
- [More information about MarianNMT models in the transformers library](https://huggingface.co/docs/transformers/model_doc/marian)
- [Tatoeba Translation Challenge](https://github.com/Helsinki-NLP/Tatoeba-Challenge/)
- [HPLT bilingual data v1 (as part of the Tatoeba Translation Challenge dataset)](https://hplt-project.org/datasets/v1)
- [A massively parallel Bible corpus](https://aclanthology.org/L14-1215/)
This is a multilingual translation model with multiple target languages. A sentence initial language token is required in the form of `>>id<<` (id = valid target language ID), e.g. `>>deu<<`
## Uses
This model can be used for translation and text-to-text generation.
## Risks, Limitations and Biases
**CONTENT WARNING: Readers should be aware that the model is trained on various public data sets that may contain content that is disturbing, offensive, and can propagate historical and current stereotypes.**
Significant research has explored bias and fairness issues with language models (see, e.g., [Sheng et al. (2021)](https://aclanthology.org/2021.acl-long.330.pdf) and [Bender et al. (2021)](https://dl.acm.org/doi/pdf/10.1145/3442188.3445922)).
## How to Get Started With the Model
A short example code:
```python
from transformers import MarianMTModel, MarianTokenizer
src_text = [
">>deu<< Replace this with text in an accepted source language.",
">>spa<< This is the second sentence."
]
model_name = "pytorch-models/opus-mt-tc-bible-big-itc-deu_eng_fra_por_spa"
tokenizer = MarianTokenizer.from_pretrained(model_name)
model = MarianMTModel.from_pretrained(model_name)
translated = model.generate(**tokenizer(src_text, return_tensors="pt", padding=True))
for t in translated:
print( tokenizer.decode(t, skip_special_tokens=True) )
```
You can also use OPUS-MT models with the transformers pipelines, for example:
```python
from transformers import pipeline
pipe = pipeline("translation", model="Helsinki-NLP/opus-mt-tc-bible-big-itc-deu_eng_fra_por_spa")
print(pipe(">>deu<< Replace this with text in an accepted source language."))
```
## Training
- **Data**: opusTCv20230926max50+bt+jhubc ([source](https://github.com/Helsinki-NLP/Tatoeba-Challenge))
- **Pre-processing**: SentencePiece (spm32k,spm32k)
- **Model Type:** transformer-big
- **Original MarianNMT Model**: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip](https://object.pouta.csc.fi/Tatoeba-MT-models/itc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30.zip)
- **Training Scripts**: [GitHub Repo](https://github.com/Helsinki-NLP/OPUS-MT-train)
## Evaluation
* [Model scores at the OPUS-MT dashboard](https://opus.nlpl.eu/dashboard/index.php?pkg=opusmt&test=all&scoreslang=all&chart=standard&model=Tatoeba-MT-models/itc-deu%2Beng%2Bfra%2Bpor%2Bspa/opusTCv20230926max50%2Bbt%2Bjhubc_transformer-big_2024-05-30)
* test set translations: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/itc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.test.txt)
* test set scores: [opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt](https://object.pouta.csc.fi/Tatoeba-MT-models/itc-deu+eng+fra+por+spa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-29.eval.txt)
* benchmark results: [benchmark_results.txt](benchmark_results.txt)
* benchmark output: [benchmark_translations.zip](benchmark_translations.zip)
| langpair | testset | chr-F | BLEU | #sent | #words |
|----------|---------|-------|-------|-------|--------|
| cat-deu | tatoeba-test-v2021-08-07 | 0.66856 | 47.9 | 723 | 5676 |
| cat-eng | tatoeba-test-v2021-08-07 | 0.72313 | 57.9 | 1631 | 12627 |
| cat-fra | tatoeba-test-v2021-08-07 | 0.71565 | 53.8 | 700 | 5664 |
| cat-por | tatoeba-test-v2021-08-07 | 0.75797 | 58.7 | 747 | 6119 |
| cat-spa | tatoeba-test-v2021-08-07 | 0.87610 | 77.7 | 1534 | 12094 |
| fra-deu | tatoeba-test-v2021-08-07 | 0.68638 | 50.0 | 12418 | 100545 |
| fra-eng | tatoeba-test-v2021-08-07 | 0.72664 | 58.0 | 12681 | 101754 |
| fra-fra | tatoeba-test-v2021-08-07 | 0.62093 | 40.6 | 1000 | 7757 |
| fra-por | tatoeba-test-v2021-08-07 | 0.70764 | 52.0 | 10518 | 77650 |
| fra-spa | tatoeba-test-v2021-08-07 | 0.72229 | 55.0 | 10294 | 78406 |
| glg-eng | tatoeba-test-v2021-08-07 | 0.70552 | 55.7 | 1015 | 8421 |
| glg-por | tatoeba-test-v2021-08-07 | 0.77067 | 62.1 | 433 | 3105 |
| glg-spa | tatoeba-test-v2021-08-07 | 0.82795 | 72.1 | 2121 | 17443 |
| ita-deu | tatoeba-test-v2021-08-07 | 0.68325 | 49.4 | 10094 | 79762 |
| ita-eng | tatoeba-test-v2021-08-07 | 0.81176 | 70.5 | 17320 | 119214 |
| ita-fra | tatoeba-test-v2021-08-07 | 0.78299 | 64.4 | 10091 | 66377 |
| ita-por | tatoeba-test-v2021-08-07 | 0.74169 | 55.6 | 3066 | 25668 |
| ita-spa | tatoeba-test-v2021-08-07 | 0.77673 | 63.0 | 5000 | 34937 |
| lad_Latn-eng | tatoeba-test-v2021-08-07 | 0.54247 | 36.7 | 672 | 3665 |
| lad_Latn-spa | tatoeba-test-v2021-08-07 | 0.59790 | 40.4 | 239 | 1239 |
| lat-deu | tatoeba-test-v2021-08-07 | 0.42548 | 24.8 | 2016 | 13326 |
| lat-eng | tatoeba-test-v2021-08-07 | 0.42385 | 24.3 | 10298 | 100152 |
| lat-spa | tatoeba-test-v2021-08-07 | 0.45821 | 25.2 | 3129 | 34036 |
| oci-eng | tatoeba-test-v2021-08-07 | 0.40921 | 22.4 | 841 | 5299 |
| oci-fra | tatoeba-test-v2021-08-07 | 0.49044 | 28.4 | 806 | 6302 |
| pcd-fra | tatoeba-test-v2021-08-07 | 0.41500 | 15.0 | 266 | 1677 |
| pms-eng | tatoeba-test-v2021-08-07 | 0.39308 | 20.8 | 269 | 2059 |
| por-deu | tatoeba-test-v2021-08-07 | 0.68379 | 48.8 | 10000 | 81246 |
| por-eng | tatoeba-test-v2021-08-07 | 0.77089 | 64.2 | 13222 | 105351 |
| por-fra | tatoeba-test-v2021-08-07 | 0.75364 | 58.7 | 10518 | 80459 |
| por-por | tatoeba-test-v2021-08-07 | 0.71396 | 50.3 | 2500 | 19220 |
| por-spa | tatoeba-test-v2021-08-07 | 0.79684 | 65.2 | 10947 | 87335 |
| ron-deu | tatoeba-test-v2021-08-07 | 0.68217 | 50.3 | 1141 | 7893 |
| ron-eng | tatoeba-test-v2021-08-07 | 0.73059 | 59.0 | 5508 | 40717 |
| ron-fra | tatoeba-test-v2021-08-07 | 0.70724 | 54.1 | 1925 | 13347 |
| ron-por | tatoeba-test-v2021-08-07 | 0.73085 | 53.3 | 681 | 4593 |
| ron-spa | tatoeba-test-v2021-08-07 | 0.73813 | 57.6 | 1959 | 12679 |
| spa-deu | tatoeba-test-v2021-08-07 | 0.68124 | 49.3 | 10521 | 86430 |
| spa-eng | tatoeba-test-v2021-08-07 | 0.74977 | 61.0 | 16583 | 138123 |
| spa-fra | tatoeba-test-v2021-08-07 | 0.73392 | 56.6 | 10294 | 83501 |
| spa-por | tatoeba-test-v2021-08-07 | 0.77280 | 61.1 | 10947 | 87610 |
| spa-spa | tatoeba-test-v2021-08-07 | 0.68111 | 50.9 | 2500 | 21469 |
| ast-deu | flores101-devtest | 0.53243 | 24.2 | 1012 | 25094 |
| ast-eng | flores101-devtest | 0.61235 | 36.0 | 1012 | 24721 |
| ast-fra | flores101-devtest | 0.56687 | 31.2 | 1012 | 28343 |
| ast-por | flores101-devtest | 0.57033 | 30.6 | 1012 | 26519 |
| ast-spa | flores101-devtest | 0.49637 | 21.2 | 1012 | 29199 |
| cat-fra | flores101-devtest | 0.63271 | 38.4 | 1012 | 28343 |
| fra-deu | flores101-devtest | 0.58433 | 28.9 | 1012 | 25094 |
| fra-eng | flores101-devtest | 0.67826 | 43.3 | 1012 | 24721 |
| glg-deu | flores101-devtest | 0.56897 | 27.1 | 1012 | 25094 |
| glg-spa | flores101-devtest | 0.53183 | 24.2 | 1012 | 29199 |
| ita-por | flores101-devtest | 0.57961 | 28.4 | 1012 | 26519 |
| kea-deu | flores101-devtest | 0.48105 | 18.3 | 1012 | 25094 |
| kea-eng | flores101-devtest | 0.60362 | 35.0 | 1012 | 24721 |
| kea-por | flores101-devtest | 0.57808 | 29.0 | 1012 | 26519 |
| kea-spa | flores101-devtest | 0.46648 | 17.6 | 1012 | 29199 |
| oci-deu | flores101-devtest | 0.57391 | 28.0 | 1012 | 25094 |
| oci-eng | flores101-devtest | 0.72351 | 49.4 | 1012 | 24721 |
| por-eng | flores101-devtest | 0.70724 | 47.4 | 1012 | 24721 |
| por-fra | flores101-devtest | 0.64103 | 39.2 | 1012 | 28343 |
| por-spa | flores101-devtest | 0.53268 | 25.0 | 1012 | 29199 |
| ron-deu | flores101-devtest | 0.57980 | 28.1 | 1012 | 25094 |
| ron-eng | flores101-devtest | 0.67583 | 41.6 | 1012 | 24721 |
| ron-spa | flores101-devtest | 0.53082 | 24.3 | 1012 | 29199 |
| spa-fra | flores101-devtest | 0.57039 | 27.1 | 1012 | 28343 |
| spa-por | flores101-devtest | 0.55607 | 25.0 | 1012 | 26519 |
| ast-deu | flores200-devtest | 0.53776 | 24.8 | 1012 | 25094 |
| ast-eng | flores200-devtest | 0.61482 | 36.8 | 1012 | 24721 |
| ast-fra | flores200-devtest | 0.56504 | 31.3 | 1012 | 28343 |
| ast-por | flores200-devtest | 0.57158 | 31.1 | 1012 | 26519 |
| ast-spa | flores200-devtest | 0.49579 | 21.2 | 1012 | 29199 |
| cat-deu | flores200-devtest | 0.58203 | 29.2 | 1012 | 25094 |
| cat-eng | flores200-devtest | 0.69165 | 44.6 | 1012 | 24721 |
| cat-fra | flores200-devtest | 0.63612 | 38.9 | 1012 | 28343 |
| cat-por | flores200-devtest | 0.62911 | 37.7 | 1012 | 26519 |
| cat-spa | flores200-devtest | 0.53320 | 24.6 | 1012 | 29199 |
| fra-deu | flores200-devtest | 0.58592 | 29.1 | 1012 | 25094 |
| fra-eng | flores200-devtest | 0.68067 | 43.8 | 1012 | 24721 |
| fra-por | flores200-devtest | 0.62388 | 37.0 | 1012 | 26519 |
| fra-spa | flores200-devtest | 0.52983 | 24.4 | 1012 | 29199 |
| fur-deu | flores200-devtest | 0.51969 | 21.8 | 1012 | 25094 |
| fur-eng | flores200-devtest | 0.60793 | 34.3 | 1012 | 24721 |
| fur-fra | flores200-devtest | 0.56989 | 30.0 | 1012 | 28343 |
| fur-por | flores200-devtest | 0.56207 | 29.3 | 1012 | 26519 |
| fur-spa | flores200-devtest | 0.48436 | 20.0 | 1012 | 29199 |
| glg-deu | flores200-devtest | 0.57369 | 27.7 | 1012 | 25094 |
| glg-eng | flores200-devtest | 0.66358 | 40.0 | 1012 | 24721 |
| glg-fra | flores200-devtest | 0.62487 | 36.5 | 1012 | 28343 |
| glg-por | flores200-devtest | 0.60267 | 32.7 | 1012 | 26519 |
| glg-spa | flores200-devtest | 0.53227 | 24.3 | 1012 | 29199 |
| hat-deu | flores200-devtest | 0.49916 | 19.1 | 1012 | 25094 |
| hat-eng | flores200-devtest | 0.59656 | 32.5 | 1012 | 24721 |
| hat-fra | flores200-devtest | 0.61574 | 35.4 | 1012 | 28343 |
| hat-por | flores200-devtest | 0.55195 | 27.7 | 1012 | 26519 |
| hat-spa | flores200-devtest | 0.47382 | 18.4 | 1012 | 29199 |
| ita-deu | flores200-devtest | 0.55779 | 24.1 | 1012 | 25094 |
| ita-eng | flores200-devtest | 0.61563 | 32.2 | 1012 | 24721 |
| ita-fra | flores200-devtest | 0.60210 | 31.2 | 1012 | 28343 |
| ita-por | flores200-devtest | 0.58279 | 28.8 | 1012 | 26519 |
| ita-spa | flores200-devtest | 0.52348 | 23.2 | 1012 | 29199 |
| kea-deu | flores200-devtest | 0.49089 | 19.3 | 1012 | 25094 |
| kea-eng | flores200-devtest | 0.60553 | 35.5 | 1012 | 24721 |
| kea-fra | flores200-devtest | 0.54027 | 26.6 | 1012 | 28343 |
| kea-por | flores200-devtest | 0.57696 | 28.9 | 1012 | 26519 |
| kea-spa | flores200-devtest | 0.46974 | 18.0 | 1012 | 29199 |
| lij-deu | flores200-devtest | 0.51695 | 22.7 | 1012 | 25094 |
| lij-eng | flores200-devtest | 0.62347 | 36.2 | 1012 | 24721 |
| lij-fra | flores200-devtest | 0.57498 | 31.4 | 1012 | 28343 |
| lij-por | flores200-devtest | 0.56183 | 29.4 | 1012 | 26519 |
| lij-spa | flores200-devtest | 0.48038 | 20.0 | 1012 | 29199 |
| lmo-deu | flores200-devtest | 0.45516 | 15.4 | 1012 | 25094 |
| lmo-eng | flores200-devtest | 0.53540 | 25.5 | 1012 | 24721 |
| lmo-fra | flores200-devtest | 0.50076 | 22.2 | 1012 | 28343 |
| lmo-por | flores200-devtest | 0.50134 | 22.9 | 1012 | 26519 |
| lmo-spa | flores200-devtest | 0.44053 | 16.2 | 1012 | 29199 |
| oci-deu | flores200-devtest | 0.57822 | 28.7 | 1012 | 25094 |
| oci-eng | flores200-devtest | 0.73030 | 50.7 | 1012 | 24721 |
| oci-fra | flores200-devtest | 0.64900 | 39.7 | 1012 | 28343 |
| oci-por | flores200-devtest | 0.63318 | 36.9 | 1012 | 26519 |
| oci-spa | flores200-devtest | 0.52269 | 22.9 | 1012 | 29199 |
| pap-deu | flores200-devtest | 0.53166 | 23.2 | 1012 | 25094 |
| pap-eng | flores200-devtest | 0.68541 | 44.6 | 1012 | 24721 |
| pap-fra | flores200-devtest | 0.57224 | 30.5 | 1012 | 28343 |
| pap-por | flores200-devtest | 0.59064 | 33.2 | 1012 | 26519 |
| pap-spa | flores200-devtest | 0.49601 | 21.7 | 1012 | 29199 |
| por-deu | flores200-devtest | 0.59047 | 30.3 | 1012 | 25094 |
| por-eng | flores200-devtest | 0.71096 | 48.0 | 1012 | 24721 |
| por-fra | flores200-devtest | 0.64555 | 40.1 | 1012 | 28343 |
| por-spa | flores200-devtest | 0.53400 | 25.1 | 1012 | 29199 |
| ron-deu | flores200-devtest | 0.58428 | 28.7 | 1012 | 25094 |
| ron-eng | flores200-devtest | 0.67719 | 41.8 | 1012 | 24721 |
| ron-fra | flores200-devtest | 0.63678 | 37.6 | 1012 | 28343 |
| ron-por | flores200-devtest | 0.62371 | 36.1 | 1012 | 26519 |
| ron-spa | flores200-devtest | 0.53150 | 24.5 | 1012 | 29199 |
| scn-deu | flores200-devtest | 0.48102 | 19.2 | 1012 | 25094 |
| scn-eng | flores200-devtest | 0.55782 | 29.6 | 1012 | 24721 |
| scn-fra | flores200-devtest | 0.52773 | 26.1 | 1012 | 28343 |
| scn-por | flores200-devtest | 0.51894 | 25.2 | 1012 | 26519 |
| scn-spa | flores200-devtest | 0.45724 | 17.9 | 1012 | 29199 |
| spa-deu | flores200-devtest | 0.53451 | 21.5 | 1012 | 25094 |
| spa-eng | flores200-devtest | 0.58896 | 28.5 | 1012 | 24721 |
| spa-fra | flores200-devtest | 0.57406 | 27.6 | 1012 | 28343 |
| spa-por | flores200-devtest | 0.55749 | 25.2 | 1012 | 26519 |
| srd-deu | flores200-devtest | 0.49238 | 19.9 | 1012 | 25094 |
| srd-eng | flores200-devtest | 0.59392 | 34.2 | 1012 | 24721 |
| srd-fra | flores200-devtest | 0.54003 | 27.6 | 1012 | 28343 |
| srd-por | flores200-devtest | 0.53842 | 27.9 | 1012 | 26519 |
| srd-spa | flores200-devtest | 0.46002 | 18.2 | 1012 | 29199 |
| vec-deu | flores200-devtest | 0.48795 | 19.3 | 1012 | 25094 |
| vec-eng | flores200-devtest | 0.56840 | 30.7 | 1012 | 24721 |
| vec-fra | flores200-devtest | 0.54164 | 27.3 | 1012 | 28343 |
| vec-por | flores200-devtest | 0.53482 | 26.2 | 1012 | 26519 |
| vec-spa | flores200-devtest | 0.46588 | 18.4 | 1012 | 29199 |
| fra-deu | generaltest2022 | 0.66476 | 42.4 | 2006 | 37696 |
| fra-deu | multi30k_test_2016_flickr | 0.61797 | 32.6 | 1000 | 12106 |
| fra-eng | multi30k_test_2016_flickr | 0.66271 | 47.2 | 1000 | 12955 |
| fra-deu | multi30k_test_2017_flickr | 0.59701 | 29.4 | 1000 | 10755 |
| fra-eng | multi30k_test_2017_flickr | 0.69422 | 50.3 | 1000 | 11374 |
| fra-deu | multi30k_test_2017_mscoco | 0.55509 | 25.7 | 461 | 5158 |
| fra-eng | multi30k_test_2017_mscoco | 0.67791 | 48.7 | 461 | 5231 |
| fra-deu | multi30k_test_2018_flickr | 0.55237 | 24.0 | 1071 | 13703 |
| fra-eng | multi30k_test_2018_flickr | 0.64722 | 43.8 | 1071 | 14689 |
| fra-eng | newsdiscusstest2015 | 0.61385 | 38.4 | 1500 | 26982 |
| fra-deu | newssyscomb2009 | 0.53530 | 23.7 | 502 | 11271 |
| fra-eng | newssyscomb2009 | 0.57297 | 31.3 | 502 | 11818 |
| fra-spa | newssyscomb2009 | 0.60233 | 34.1 | 502 | 12503 |
| ita-deu | newssyscomb2009 | 0.53590 | 22.4 | 502 | 11271 |
| ita-eng | newssyscomb2009 | 0.59976 | 34.8 | 502 | 11818 |
| ita-fra | newssyscomb2009 | 0.61232 | 33.5 | 502 | 12331 |
| ita-spa | newssyscomb2009 | 0.60782 | 35.3 | 502 | 12503 |
| spa-deu | newssyscomb2009 | 0.52853 | 21.8 | 502 | 11271 |
| spa-eng | newssyscomb2009 | 0.57347 | 31.0 | 502 | 11818 |
| spa-fra | newssyscomb2009 | 0.61436 | 34.3 | 502 | 12331 |
| fra-deu | newstest2008 | 0.53180 | 22.9 | 2051 | 47447 |
| fra-eng | newstest2008 | 0.54379 | 26.5 | 2051 | 49380 |
| fra-spa | newstest2008 | 0.58804 | 33.1 | 2051 | 52586 |
| spa-deu | newstest2008 | 0.52221 | 21.6 | 2051 | 47447 |
| spa-eng | newstest2008 | 0.55331 | 27.9 | 2051 | 49380 |
| spa-fra | newstest2008 | 0.58769 | 32.0 | 2051 | 52685 |
| fra-deu | newstest2009 | 0.52771 | 22.5 | 2525 | 62816 |
| fra-eng | newstest2009 | 0.56679 | 30.2 | 2525 | 65399 |
| fra-spa | newstest2009 | 0.58921 | 32.1 | 2525 | 68111 |
| ita-deu | newstest2009 | 0.53022 | 22.8 | 2525 | 62816 |
| ita-eng | newstest2009 | 0.59309 | 33.8 | 2525 | 65399 |
| ita-fra | newstest2009 | 0.59309 | 32.0 | 2525 | 69263 |
| ita-spa | newstest2009 | 0.59760 | 33.5 | 2525 | 68111 |
| spa-deu | newstest2009 | 0.52822 | 22.3 | 2525 | 62816 |
| spa-eng | newstest2009 | 0.56989 | 30.4 | 2525 | 65399 |
| spa-fra | newstest2009 | 0.59150 | 32.2 | 2525 | 69263 |
| fra-deu | newstest2010 | 0.53765 | 24.0 | 2489 | 61503 |
| fra-eng | newstest2010 | 0.59251 | 32.6 | 2489 | 61711 |
| fra-spa | newstest2010 | 0.62480 | 37.6 | 2489 | 65480 |
| spa-deu | newstest2010 | 0.55161 | 26.0 | 2489 | 61503 |
| spa-eng | newstest2010 | 0.61562 | 36.3 | 2489 | 61711 |
| spa-fra | newstest2010 | 0.62021 | 35.7 | 2489 | 66022 |
| fra-deu | newstest2011 | 0.53025 | 23.1 | 3003 | 72981 |
| fra-eng | newstest2011 | 0.59636 | 32.9 | 3003 | 74681 |
| fra-spa | newstest2011 | 0.63203 | 39.9 | 3003 | 79476 |
| spa-deu | newstest2011 | 0.52934 | 23.3 | 3003 | 72981 |
| spa-eng | newstest2011 | 0.59606 | 33.8 | 3003 | 74681 |
| spa-fra | newstest2011 | 0.61079 | 34.9 | 3003 | 80626 |
| fra-deu | newstest2012 | 0.52957 | 24.0 | 3003 | 72886 |
| fra-eng | newstest2012 | 0.59352 | 33.6 | 3003 | 72812 |
| fra-spa | newstest2012 | 0.62641 | 39.2 | 3003 | 79006 |
| spa-deu | newstest2012 | 0.53519 | 24.6 | 3003 | 72886 |
| spa-eng | newstest2012 | 0.62284 | 37.4 | 3003 | 72812 |
| spa-fra | newstest2012 | 0.61076 | 33.8 | 3003 | 78011 |
| fra-deu | newstest2013 | 0.54167 | 25.4 | 3000 | 63737 |
| fra-eng | newstest2013 | 0.59236 | 34.0 | 3000 | 64505 |
| fra-spa | newstest2013 | 0.59347 | 34.9 | 3000 | 70528 |
| spa-deu | newstest2013 | 0.55130 | 26.3 | 3000 | 63737 |
| spa-eng | newstest2013 | 0.60681 | 34.6 | 3000 | 64505 |
| spa-fra | newstest2013 | 0.59816 | 33.2 | 3000 | 70037 |
| fra-eng | newstest2014 | 0.63499 | 37.9 | 3003 | 70708 |
| ron-eng | newstest2016 | 0.63996 | 39.5 | 1999 | 47562 |
| fra-deu | newstest2019 | 0.60468 | 28.6 | 1701 | 36446 |
| fra-deu | newstest2020 | 0.61401 | 28.8 | 1619 | 30265 |
| fra-deu | newstest2021 | 0.65950 | 39.5 | 1026 | 26077 |
| cat-deu | ntrex128 | 0.54096 | 24.0 | 1997 | 48761 |
| cat-eng | ntrex128 | 0.63516 | 36.5 | 1997 | 47673 |
| cat-fra | ntrex128 | 0.56385 | 28.1 | 1997 | 53481 |
| cat-por | ntrex128 | 0.56246 | 28.7 | 1997 | 51631 |
| cat-spa | ntrex128 | 0.61311 | 35.8 | 1997 | 54107 |
| fra-deu | ntrex128 | 0.53059 | 23.4 | 1997 | 48761 |
| fra-eng | ntrex128 | 0.61285 | 34.7 | 1997 | 47673 |
| fra-por | ntrex128 | 0.54075 | 25.8 | 1997 | 51631 |
| fra-spa | ntrex128 | 0.56863 | 30.6 | 1997 | 54107 |
| glg-deu | ntrex128 | 0.53724 | 23.6 | 1997 | 48761 |
| glg-eng | ntrex128 | 0.64481 | 38.7 | 1997 | 47673 |
| glg-fra | ntrex128 | 0.55856 | 27.8 | 1997 | 53481 |
| glg-por | ntrex128 | 0.56322 | 28.7 | 1997 | 51631 |
| glg-spa | ntrex128 | 0.61794 | 36.8 | 1997 | 54107 |
| ita-deu | ntrex128 | 0.54678 | 25.0 | 1997 | 48761 |
| ita-eng | ntrex128 | 0.64636 | 39.2 | 1997 | 47673 |
| ita-fra | ntrex128 | 0.57428 | 30.0 | 1997 | 53481 |
| ita-por | ntrex128 | 0.56858 | 29.7 | 1997 | 51631 |
| ita-spa | ntrex128 | 0.58886 | 33.0 | 1997 | 54107 |
| por-deu | ntrex128 | 0.54833 | 24.6 | 1997 | 48761 |
| por-eng | ntrex128 | 0.65223 | 39.7 | 1997 | 47673 |
| por-fra | ntrex128 | 0.56793 | 28.9 | 1997 | 53481 |
| por-spa | ntrex128 | 0.59218 | 33.8 | 1997 | 54107 |
| ron-deu | ntrex128 | 0.53249 | 22.4 | 1997 | 48761 |
| ron-eng | ntrex128 | 0.61807 | 33.8 | 1997 | 47673 |
| ron-fra | ntrex128 | 0.55575 | 26.4 | 1997 | 53481 |
| ron-por | ntrex128 | 0.55086 | 27.2 | 1997 | 51631 |
| ron-spa | ntrex128 | 0.57787 | 31.9 | 1997 | 54107 |
| spa-deu | ntrex128 | 0.54309 | 23.8 | 1997 | 48761 |
| spa-eng | ntrex128 | 0.64416 | 37.4 | 1997 | 47673 |
| spa-fra | ntrex128 | 0.57320 | 29.4 | 1997 | 53481 |
| spa-por | ntrex128 | 0.56751 | 29.0 | 1997 | 51631 |
| fra-eng | tico19-test | 0.62364 | 39.7 | 2100 | 56323 |
| fra-por | tico19-test | 0.58563 | 34.2 | 2100 | 62729 |
| fra-spa | tico19-test | 0.59556 | 36.5 | 2100 | 66563 |
| por-eng | tico19-test | 0.74420 | 51.8 | 2100 | 56315 |
| por-fra | tico19-test | 0.60081 | 34.5 | 2100 | 64661 |
| por-spa | tico19-test | 0.68156 | 44.8 | 2100 | 66563 |
| spa-eng | tico19-test | 0.73454 | 50.3 | 2100 | 56315 |
| spa-fra | tico19-test | 0.60441 | 34.9 | 2100 | 64661 |
| spa-por | tico19-test | 0.67749 | 42.7 | 2100 | 62729 |
## Citation Information
* Publications: [Democratizing neural machine translation with OPUS-MT](https://doi.org/10.1007/s10579-023-09704-w) and [OPUS-MT – Building open translation services for the World](https://aclanthology.org/2020.eamt-1.61/) and [The Tatoeba Translation Challenge – Realistic Data Sets for Low Resource and Multilingual MT](https://aclanthology.org/2020.wmt-1.139/) (Please, cite if you use this model.)
```bibtex
@article{tiedemann2023democratizing,
title={Democratizing neural machine translation with {OPUS-MT}},
author={Tiedemann, J{\"o}rg and Aulamo, Mikko and Bakshandaeva, Daria and Boggia, Michele and Gr{\"o}nroos, Stig-Arne and Nieminen, Tommi and Raganato, Alessandro and Scherrer, Yves and Vazquez, Raul and Virpioja, Sami},
journal={Language Resources and Evaluation},
number={58},
pages={713--755},
year={2023},
publisher={Springer Nature},
issn={1574-0218},
doi={10.1007/s10579-023-09704-w}
}
@inproceedings{tiedemann-thottingal-2020-opus,
title = "{OPUS}-{MT} {--} Building open translation services for the World",
author = {Tiedemann, J{\"o}rg and Thottingal, Santhosh},
booktitle = "Proceedings of the 22nd Annual Conference of the European Association for Machine Translation",
month = nov,
year = "2020",
address = "Lisboa, Portugal",
publisher = "European Association for Machine Translation",
url = "https://aclanthology.org/2020.eamt-1.61",
pages = "479--480",
}
@inproceedings{tiedemann-2020-tatoeba,
title = "The Tatoeba Translation Challenge {--} Realistic Data Sets for Low Resource and Multilingual {MT}",
author = {Tiedemann, J{\"o}rg},
booktitle = "Proceedings of the Fifth Conference on Machine Translation",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2020.wmt-1.139",
pages = "1174--1182",
}
```
## Acknowledgements
The work is supported by the [HPLT project](https://hplt-project.org/), funded by the European Union’s Horizon Europe research and innovation programme under grant agreement No 101070350. We are also grateful for the generous computational resources and IT infrastructure provided by [CSC -- IT Center for Science](https://www.csc.fi/), Finland, and the [EuroHPC supercomputer LUMI](https://www.lumi-supercomputer.eu/).
## Model conversion info
* transformers version: 4.45.1
* OPUS-MT git hash: 0882077
* port time: Tue Oct 8 11:57:19 EEST 2024
* port machine: LM0-400-22516.local
| [
"TRANSLATION"
] | Non_BioNLP |
gokuls/BERT-tiny-Massive-intent | gokuls | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:massive",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,664,028,930,000 | 2022-09-24T14:26:13 | 10 | 0 | ---
datasets:
- massive
license: apache-2.0
metrics:
- accuracy
tags:
- generated_from_trainer
model-index:
- name: BERT-tiny-Massive-intent
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: massive
type: massive
config: en-US
split: train
args: en-US
metrics:
- type: accuracy
value: 0.8475159862272503
name: Accuracy
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# BERT-tiny-Massive-intent
This model is a fine-tuned version of [google/bert_uncased_L-2_H-128_A-2](https://huggingface.co/google/bert_uncased_L-2_H-128_A-2) on the massive dataset.
It achieves the following results on the evaluation set:
- Loss: 0.6740
- Accuracy: 0.8475
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 33
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 3.6104 | 1.0 | 720 | 3.0911 | 0.3601 |
| 2.8025 | 2.0 | 1440 | 2.3800 | 0.5165 |
| 2.2292 | 3.0 | 2160 | 1.9134 | 0.5991 |
| 1.818 | 4.0 | 2880 | 1.5810 | 0.6744 |
| 1.5171 | 5.0 | 3600 | 1.3522 | 0.7108 |
| 1.2876 | 6.0 | 4320 | 1.1686 | 0.7442 |
| 1.1049 | 7.0 | 5040 | 1.0355 | 0.7683 |
| 0.9623 | 8.0 | 5760 | 0.9466 | 0.7885 |
| 0.8424 | 9.0 | 6480 | 0.8718 | 0.7875 |
| 0.7473 | 10.0 | 7200 | 0.8107 | 0.8028 |
| 0.6735 | 11.0 | 7920 | 0.7710 | 0.8180 |
| 0.6085 | 12.0 | 8640 | 0.7404 | 0.8210 |
| 0.5536 | 13.0 | 9360 | 0.7180 | 0.8229 |
| 0.5026 | 14.0 | 10080 | 0.6980 | 0.8318 |
| 0.4652 | 15.0 | 10800 | 0.6970 | 0.8337 |
| 0.4234 | 16.0 | 11520 | 0.6822 | 0.8372 |
| 0.3987 | 17.0 | 12240 | 0.6691 | 0.8436 |
| 0.3707 | 18.0 | 12960 | 0.6679 | 0.8455 |
| 0.3433 | 19.0 | 13680 | 0.6740 | 0.8475 |
| 0.3206 | 20.0 | 14400 | 0.6760 | 0.8451 |
| 0.308 | 21.0 | 15120 | 0.6704 | 0.8436 |
| 0.2813 | 22.0 | 15840 | 0.6701 | 0.8416 |
### Framework versions
- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.5.1
- Tokenizers 0.12.1
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
osanseviero/bert-base-uncased-copy | osanseviero | fill-mask | [
"transformers",
"pytorch",
"jax",
"rust",
"coreml",
"safetensors",
"bert",
"fill-mask",
"exbert",
"en",
"dataset:bookcorpus",
"dataset:wikipedia",
"arxiv:1810.04805",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,680,204,589,000 | 2023-04-04T06:18:11 | 14 | 0 | ---
datasets:
- bookcorpus
- wikipedia
language: en
license: apache-2.0
tags:
- exbert
duplicated_from: bert-base-uncased
---
# BERT base model (uncased)
Pretrained model on English language using a masked language modeling (MLM) objective. It was introduced in
[this paper](https://arxiv.org/abs/1810.04805) and first released in
[this repository](https://github.com/google-research/bert). This model is uncased: it does not make a difference
between english and English.
Disclaimer: The team releasing BERT did not write a model card for this model so this model card has been written by
the Hugging Face team.
## Model description
BERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it
was pretrained on the raw texts only, with no humans labeling them in any way (which is why it can use lots of
publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it
was pretrained with two objectives:
- Masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then run
the entire masked sentence through the model and has to predict the masked words. This is different from traditional
recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like
GPT which internally masks the future tokens. It allows the model to learn a bidirectional representation of the
sentence.
- Next sentence prediction (NSP): the models concatenates two masked sentences as inputs during pretraining. Sometimes
they correspond to sentences that were next to each other in the original text, sometimes not. The model then has to
predict if the two sentences were following each other or not.
This way, the model learns an inner representation of the English language that can then be used to extract features
useful for downstream tasks: if you have a dataset of labeled sentences, for instance, you can train a standard
classifier using the features produced by the BERT model as inputs.
## Model variations
BERT has originally been released in base and large variations, for cased and uncased input text. The uncased models also strips out an accent markers.
Chinese and multilingual uncased and cased versions followed shortly after.
Modified preprocessing with whole word masking has replaced subpiece masking in a following work, with the release of two models.
Other 24 smaller models are released afterward.
The detailed release history can be found on the [google-research/bert readme](https://github.com/google-research/bert/blob/master/README.md) on github.
| Model | #params | Language |
|------------------------|--------------------------------|-------|
| [`bert-base-uncased`](https://huggingface.co/bert-base-uncased) | 110M | English |
| [`bert-large-uncased`](https://huggingface.co/bert-large-uncased) | 340M | English | sub
| [`bert-base-cased`](https://huggingface.co/bert-base-cased) | 110M | English |
| [`bert-large-cased`](https://huggingface.co/bert-large-cased) | 340M | English |
| [`bert-base-chinese`](https://huggingface.co/bert-base-chinese) | 110M | Chinese |
| [`bert-base-multilingual-cased`](https://huggingface.co/bert-base-multilingual-cased) | 110M | Multiple |
| [`bert-large-uncased-whole-word-masking`](https://huggingface.co/bert-large-uncased-whole-word-masking) | 340M | English |
| [`bert-large-cased-whole-word-masking`](https://huggingface.co/bert-large-cased-whole-word-masking) | 340M | English |
## Intended uses & limitations
You can use the raw model for either masked language modeling or next sentence prediction, but it's mostly intended to
be fine-tuned on a downstream task. See the [model hub](https://huggingface.co/models?filter=bert) to look for
fine-tuned versions of a task that interests you.
Note that this model is primarily aimed at being fine-tuned on tasks that use the whole sentence (potentially masked)
to make decisions, such as sequence classification, token classification or question answering. For tasks such as text
generation you should look at model like GPT2.
### How to use
You can use this model directly with a pipeline for masked language modeling:
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='bert-base-uncased')
>>> unmasker("Hello I'm a [MASK] model.")
[{'sequence': "[CLS] hello i'm a fashion model. [SEP]",
'score': 0.1073106899857521,
'token': 4827,
'token_str': 'fashion'},
{'sequence': "[CLS] hello i'm a role model. [SEP]",
'score': 0.08774490654468536,
'token': 2535,
'token_str': 'role'},
{'sequence': "[CLS] hello i'm a new model. [SEP]",
'score': 0.05338378623127937,
'token': 2047,
'token_str': 'new'},
{'sequence': "[CLS] hello i'm a super model. [SEP]",
'score': 0.04667217284440994,
'token': 3565,
'token_str': 'super'},
{'sequence': "[CLS] hello i'm a fine model. [SEP]",
'score': 0.027095865458250046,
'token': 2986,
'token_str': 'fine'}]
```
Here is how to use this model to get the features of a given text in PyTorch:
```python
from transformers import BertTokenizer, BertModel
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
model = BertModel.from_pretrained("bert-base-uncased")
text = "Replace me by any text you'd like."
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
```
and in TensorFlow:
```python
from transformers import BertTokenizer, TFBertModel
tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
model = TFBertModel.from_pretrained("bert-base-uncased")
text = "Replace me by any text you'd like."
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
```
### Limitations and bias
Even if the training data used for this model could be characterized as fairly neutral, this model can have biased
predictions:
```python
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='bert-base-uncased')
>>> unmasker("The man worked as a [MASK].")
[{'sequence': '[CLS] the man worked as a carpenter. [SEP]',
'score': 0.09747550636529922,
'token': 10533,
'token_str': 'carpenter'},
{'sequence': '[CLS] the man worked as a waiter. [SEP]',
'score': 0.0523831807076931,
'token': 15610,
'token_str': 'waiter'},
{'sequence': '[CLS] the man worked as a barber. [SEP]',
'score': 0.04962705448269844,
'token': 13362,
'token_str': 'barber'},
{'sequence': '[CLS] the man worked as a mechanic. [SEP]',
'score': 0.03788609802722931,
'token': 15893,
'token_str': 'mechanic'},
{'sequence': '[CLS] the man worked as a salesman. [SEP]',
'score': 0.037680890411138535,
'token': 18968,
'token_str': 'salesman'}]
>>> unmasker("The woman worked as a [MASK].")
[{'sequence': '[CLS] the woman worked as a nurse. [SEP]',
'score': 0.21981462836265564,
'token': 6821,
'token_str': 'nurse'},
{'sequence': '[CLS] the woman worked as a waitress. [SEP]',
'score': 0.1597415804862976,
'token': 13877,
'token_str': 'waitress'},
{'sequence': '[CLS] the woman worked as a maid. [SEP]',
'score': 0.1154729500412941,
'token': 10850,
'token_str': 'maid'},
{'sequence': '[CLS] the woman worked as a prostitute. [SEP]',
'score': 0.037968918681144714,
'token': 19215,
'token_str': 'prostitute'},
{'sequence': '[CLS] the woman worked as a cook. [SEP]',
'score': 0.03042375110089779,
'token': 5660,
'token_str': 'cook'}]
```
This bias will also affect all fine-tuned versions of this model.
## Training data
The BERT model was pretrained on [BookCorpus](https://yknzhu.wixsite.com/mbweb), a dataset consisting of 11,038
unpublished books and [English Wikipedia](https://en.wikipedia.org/wiki/English_Wikipedia) (excluding lists, tables and
headers).
## Training procedure
### Preprocessing
The texts are lowercased and tokenized using WordPiece and a vocabulary size of 30,000. The inputs of the model are
then of the form:
```
[CLS] Sentence A [SEP] Sentence B [SEP]
```
With probability 0.5, sentence A and sentence B correspond to two consecutive sentences in the original corpus, and in
the other cases, it's another random sentence in the corpus. Note that what is considered a sentence here is a
consecutive span of text usually longer than a single sentence. The only constrain is that the result with the two
"sentences" has a combined length of less than 512 tokens.
The details of the masking procedure for each sentence are the following:
- 15% of the tokens are masked.
- In 80% of the cases, the masked tokens are replaced by `[MASK]`.
- In 10% of the cases, the masked tokens are replaced by a random token (different) from the one they replace.
- In the 10% remaining cases, the masked tokens are left as is.
### Pretraining
The model was trained on 4 cloud TPUs in Pod configuration (16 TPU chips total) for one million steps with a batch size
of 256. The sequence length was limited to 128 tokens for 90% of the steps and 512 for the remaining 10%. The optimizer
used is Adam with a learning rate of 1e-4, \\(\beta_{1} = 0.9\\) and \\(\beta_{2} = 0.999\\), a weight decay of 0.01,
learning rate warmup for 10,000 steps and linear decay of the learning rate after.
## Evaluation results
When fine-tuned on downstream tasks, this model achieves the following results:
Glue test results:
| Task | MNLI-(m/mm) | QQP | QNLI | SST-2 | CoLA | STS-B | MRPC | RTE | Average |
|:----:|:-----------:|:----:|:----:|:-----:|:----:|:-----:|:----:|:----:|:-------:|
| | 84.6/83.4 | 71.2 | 90.5 | 93.5 | 52.1 | 85.8 | 88.9 | 66.4 | 79.6 |
### BibTeX entry and citation info
```bibtex
@article{DBLP:journals/corr/abs-1810-04805,
author = {Jacob Devlin and
Ming{-}Wei Chang and
Kenton Lee and
Kristina Toutanova},
title = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language
Understanding},
journal = {CoRR},
volume = {abs/1810.04805},
year = {2018},
url = {http://arxiv.org/abs/1810.04805},
archivePrefix = {arXiv},
eprint = {1810.04805},
timestamp = {Tue, 30 Oct 2018 20:39:56 +0100},
biburl = {https://dblp.org/rec/journals/corr/abs-1810-04805.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
```
<a href="https://huggingface.co/exbert/?model=bert-base-uncased">
<img width="300px" src="https://cdn-media.huggingface.co/exbert/button.png">
</a>
| [
"QUESTION_ANSWERING"
] | Non_BioNLP |
sud977/my-awesome-setfit-model | sud977 | text-classification | [
"sentence-transformers",
"pytorch",
"mpnet",
"setfit",
"text-classification",
"arxiv:2209.11055",
"license:apache-2.0",
"region:us"
] | 1,682,560,241,000 | 2023-04-27T01:53:28 | 9 | 0 | ---
license: apache-2.0
pipeline_tag: text-classification
tags:
- setfit
- sentence-transformers
- text-classification
---
# /var/folders/lm/k69sycyx5538ldsk5n0ln5000000gn/T/tmp_un7plj_/killshot977/my-awesome-setfit-model
This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves:
1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.
## Usage
To use this model for inference, first install the SetFit library:
```bash
python -m pip install setfit
```
You can then run inference as follows:
```python
from setfit import SetFitModel
# Download from Hub and run inference
model = SetFitModel.from_pretrained("/var/folders/lm/k69sycyx5538ldsk5n0ln5000000gn/T/tmp_un7plj_/killshot977/my-awesome-setfit-model")
# Run inference
preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"])
```
## BibTeX entry and citation info
```bibtex
@article{https://doi.org/10.48550/arxiv.2209.11055,
doi = {10.48550/ARXIV.2209.11055},
url = {https://arxiv.org/abs/2209.11055},
author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Efficient Few-Shot Learning Without Prompts},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}
}
```
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
YtBig/improve-a-v1 | YtBig | summarization | [
"transformers",
"pytorch",
"bart",
"text2text-generation",
"autotrain",
"summarization",
"en",
"co2_eq_emissions",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,666,707,054,000 | 2022-12-08T09:13:15 | 114 | 0 | ---
language:
- en
tags:
- autotrain
- summarization
widget:
- text: I love AutoTrain 🤗
co2_eq_emissions:
emissions: 0.9899872350262614
---
# Model Trained Using AutoTrain
- Problem type: Summarization
- Model ID: 1822063032
- CO2 Emissions (in grams): 0.9900
## Validation Metrics
- Loss: 0.347
- Rouge1: 66.429
- Rouge2: 29.419
- RougeL: 66.188
- RougeLsum: 66.183
- Gen Len: 11.256
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_HUGGINGFACE_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoTrain"}' https://api-inference.huggingface.co/aalbertini90/autotrain-improve-a-1822063032
``` | [
"SUMMARIZATION"
] | Non_BioNLP |
gokuls/bert_uncased_L-2_H-768_A-12_massive | gokuls | text-classification | [
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"dataset:massive",
"base_model:google/bert_uncased_L-2_H-768_A-12",
"base_model:finetune:google/bert_uncased_L-2_H-768_A-12",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,696,613,565,000 | 2023-10-06T17:35:40 | 10 | 0 | ---
base_model: google/bert_uncased_L-2_H-768_A-12
datasets:
- massive
license: apache-2.0
metrics:
- accuracy
tags:
- generated_from_trainer
model-index:
- name: bert_uncased_L-2_H-768_A-12_massive
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: massive
type: massive
config: en-US
split: validation
args: en-US
metrics:
- type: accuracy
value: 0.8745696015740285
name: Accuracy
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bert_uncased_L-2_H-768_A-12_massive
This model is a fine-tuned version of [google/bert_uncased_L-2_H-768_A-12](https://huggingface.co/google/bert_uncased_L-2_H-768_A-12) on the massive dataset.
It achieves the following results on the evaluation set:
- Loss: 0.5434
- Accuracy: 0.8746
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 33
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 2.5143 | 1.0 | 180 | 1.2564 | 0.7024 |
| 1.0135 | 2.0 | 360 | 0.7279 | 0.8205 |
| 0.6173 | 3.0 | 540 | 0.5817 | 0.8559 |
| 0.433 | 4.0 | 720 | 0.5234 | 0.8598 |
| 0.312 | 5.0 | 900 | 0.5019 | 0.8657 |
| 0.23 | 6.0 | 1080 | 0.5028 | 0.8711 |
| 0.1742 | 7.0 | 1260 | 0.5037 | 0.8682 |
| 0.1314 | 8.0 | 1440 | 0.5018 | 0.8692 |
| 0.1031 | 9.0 | 1620 | 0.5188 | 0.8731 |
| 0.081 | 10.0 | 1800 | 0.5231 | 0.8711 |
| 0.0671 | 11.0 | 1980 | 0.5407 | 0.8716 |
| 0.0569 | 12.0 | 2160 | 0.5309 | 0.8721 |
| 0.0466 | 13.0 | 2340 | 0.5463 | 0.8711 |
| 0.0414 | 14.0 | 2520 | 0.5434 | 0.8746 |
| 0.039 | 15.0 | 2700 | 0.5464 | 0.8721 |
### Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
nielsr/coref-bert-large | nielsr | null | [
"transformers",
"pytorch",
"safetensors",
"exbert",
"en",
"dataset:wikipedia",
"dataset:quoref",
"dataset:docred",
"dataset:fever",
"dataset:gap",
"dataset:winograd_wsc",
"dataset:winogender",
"dataset:nyu-mll/glue",
"arxiv:2004.06870",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | 1,646,263,745,000 | 2024-12-22T10:40:56 | 38 | 1 | ---
datasets:
- wikipedia
- quoref
- docred
- fever
- gap
- winograd_wsc
- winogender
- nyu-mll/glue
language: en
license: apache-2.0
tags:
- exbert
---
# CorefBERT large model
Pretrained model on English language using Masked Language Modeling (MLM) and Mention Reference Prediction (MRP) objectives. It was introduced in
[this paper](https://arxiv.org/abs/2004.06870) and first released in
[this repository](https://github.com/thunlp/CorefBERT).
Disclaimer: The team releasing CorefBERT did not write a model card for this model so this model card has been written by me.
## Model description
CorefBERT is a transformers model pretrained on a large corpus of English data in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of
publicly available data) with an automatic process to generate inputs and labels from those texts. More precisely, it was pretrained with two objectives:
- Masked language modeling (MLM): taking a sentence, the model randomly masks 15% of the words in the input then run
the entire masked sentence through the model and has to predict the masked words. This is different from traditional
recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like
GPT which internally mask the future tokens. It allows the model to learn a bidirectional representation of the
sentence.
- Mention reference prediction (MRP): this is a novel training task which is proposed to enhance coreferential reasoning ability. MRP utilizes the
mention reference masking strategy to mask one of the repeated mentions and then employs a copybased training objective to predict the masked tokens by copying from other tokens in the sequence.
This way, the model learns an inner representation of the English language that can then be used to extract features useful for downstream tasks, especially those that involve coreference resolution. If you have a dataset of labeled sentences for instance, you can train a standard classifier using the features produced by the CorefBERT model as inputs.
### BibTeX entry and citation info
```bibtex
@misc{ye2020coreferential,
title={Coreferential Reasoning Learning for Language Representation},
author={Deming Ye and Yankai Lin and Jiaju Du and Zhenghao Liu and Peng Li and Maosong Sun and Zhiyuan Liu},
year={2020},
eprint={2004.06870},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
| [
"COREFERENCE_RESOLUTION"
] | Non_BioNLP |
jjae/kobart-summarization-diary | jjae | text2text-generation | [
"transformers",
"safetensors",
"bart",
"text2text-generation",
"kobart-summarization-diary",
"generated_from_trainer",
"base_model:gogamza/kobart-summarization",
"base_model:finetune:gogamza/kobart-summarization",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,712,159,644,000 | 2024-04-03T16:46:18 | 14 | 0 | ---
base_model: gogamza/kobart-summarization
license: mit
tags:
- kobart-summarization-diary
- generated_from_trainer
model-index:
- name: summary2
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# summary2
This model is a fine-tuned version of [gogamza/kobart-summarization](https://huggingface.co/gogamza/kobart-summarization) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3377
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5.6e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 300
- num_epochs: 50
### Training results
| Training Loss | Epoch | Step | Validation Loss |
|:-------------:|:-----:|:----:|:---------------:|
| 1.5089 | 1.23 | 500 | 0.3360 |
| 0.238 | 2.47 | 1000 | 0.3377 |
| 0.1456 | 3.7 | 1500 | 0.3513 |
| 0.0848 | 4.94 | 2000 | 0.3753 |
| 0.0482 | 6.17 | 2500 | 0.4024 |
### Framework versions
- Transformers 4.39.3
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.0
| [
"SUMMARIZATION"
] | Non_BioNLP |
tomaarsen/distilroberta-base-nli-v2 | tomaarsen | sentence-similarity | [
"sentence-transformers",
"safetensors",
"roberta",
"sentence-similarity",
"feature-extraction",
"loss:MultipleNegativesRankingLoss",
"en",
"arxiv:1908.10084",
"arxiv:1705.00652",
"base_model:distilbert/distilroberta-base",
"base_model:finetune:distilbert/distilroberta-base",
"model-index",
"co2_eq_emissions",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,714,636,181,000 | 2024-05-02T07:50:07 | 9 | 0 | ---
base_model: distilbert/distilroberta-base
language:
- en
library_name: sentence-transformers
metrics:
- pearson_cosine
- spearman_cosine
- pearson_manhattan
- spearman_manhattan
- pearson_euclidean
- spearman_euclidean
- pearson_dot
- spearman_dot
- pearson_max
- spearman_max
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- loss:MultipleNegativesRankingLoss
widget:
- source_sentence: There's a dock
sentences:
- A boat docked on a river.
- The girl is standing.
- The boy is sleeping.
- source_sentence: The boy scowls
sentences:
- The boy is smiling
- A story book is open.
- Two women are sleeping.
- source_sentence: A bird flying.
sentences:
- an eagle flies
- The person is amused.
- Two men are sleeping.
- source_sentence: an eagle flies
sentences:
- A butterfly flys freely.
- Two men are sleeping.
- Some men sleep.
- source_sentence: A woman sings.
sentences:
- The woman is singing.
- a man is wearing blue
- The boy is sleeping.
co2_eq_emissions:
emissions: 1.414068558007261
energy_consumed: 0.003637924574628535
source: codecarbon
training_type: fine-tuning
on_cloud: false
cpu_model: 13th Gen Intel(R) Core(TM) i7-13700K
ram_total_size: 31.777088165283203
hours_used: 0.02
hardware_used: 1 x NVIDIA GeForce RTX 3090
model-index:
- name: SentenceTransformer based on distilbert/distilroberta-base
results:
- task:
type: semantic-similarity
name: Semantic Similarity
dataset:
name: sts dev
type: sts-dev
metrics:
- type: pearson_cosine
value: 0.7472500570689873
name: Pearson Cosine
- type: spearman_cosine
value: 0.7815286852337371
name: Spearman Cosine
- type: pearson_manhattan
value: 0.7466164303556344
name: Pearson Manhattan
- type: spearman_manhattan
value: 0.7564406124153681
name: Spearman Manhattan
- type: pearson_euclidean
value: 0.7470476982963574
name: Pearson Euclidean
- type: spearman_euclidean
value: 0.7553538112024218
name: Spearman Euclidean
- type: pearson_dot
value: 0.46791742113291
name: Pearson Dot
- type: spearman_dot
value: 0.48306144010812363
name: Spearman Dot
- type: pearson_max
value: 0.7472500570689873
name: Pearson Max
- type: spearman_max
value: 0.7815286852337371
name: Spearman Max
- task:
type: semantic-similarity
name: Semantic Similarity
dataset:
name: sts test
type: sts-test
metrics:
- type: pearson_cosine
value: 0.7145936155377322
name: Pearson Cosine
- type: spearman_cosine
value: 0.7188509446042572
name: Spearman Cosine
- type: pearson_manhattan
value: 0.7144637059488601
name: Pearson Manhattan
- type: spearman_manhattan
value: 0.7051742909657058
name: Spearman Manhattan
- type: pearson_euclidean
value: 0.7150126984629757
name: Pearson Euclidean
- type: spearman_euclidean
value: 0.7054604043597239
name: Spearman Euclidean
- type: pearson_dot
value: 0.4317482386066799
name: Pearson Dot
- type: spearman_dot
value: 0.4292906929274994
name: Spearman Dot
- type: pearson_max
value: 0.7150126984629757
name: Pearson Max
- type: spearman_max
value: 0.7188509446042572
name: Spearman Max
---
# SentenceTransformer based on distilbert/distilroberta-base
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [distilbert/distilroberta-base](https://huggingface.co/distilbert/distilroberta-base) on the [sentence-transformers/all-nli](https://huggingface.co/datasets/sentence-transformers/all-nli) dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [distilbert/distilroberta-base](https://huggingface.co/distilbert/distilroberta-base) <!-- at revision fb53ab8802853c8e4fbdbcd0529f21fc6f459b2b -->
- **Maximum Sequence Length:** 512 tokens
- **Output Dimensionality:** 768 tokens
- **Similarity Function:** Cosine Similarity
- **Training Dataset:**
- [sentence-transformers/all-nli](https://huggingface.co/datasets/sentence-transformers/all-nli)
- **Language:** en
<!-- - **License:** Unknown -->
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: RobertaModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("tomaarsen/distilroberta-base-nli-v2")
# Run inference
sentences = [
'A woman sings.',
'The woman is singing.',
'a man is wearing blue',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
## Evaluation
### Metrics
#### Semantic Similarity
* Dataset: `sts-dev`
* Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| pearson_cosine | 0.7473 |
| **spearman_cosine** | **0.7815** |
| pearson_manhattan | 0.7466 |
| spearman_manhattan | 0.7564 |
| pearson_euclidean | 0.747 |
| spearman_euclidean | 0.7554 |
| pearson_dot | 0.4679 |
| spearman_dot | 0.4831 |
| pearson_max | 0.7473 |
| spearman_max | 0.7815 |
#### Semantic Similarity
* Dataset: `sts-test`
* Evaluated with [<code>EmbeddingSimilarityEvaluator</code>](https://sbert.net/docs/package_reference/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| pearson_cosine | 0.7146 |
| **spearman_cosine** | **0.7189** |
| pearson_manhattan | 0.7145 |
| spearman_manhattan | 0.7052 |
| pearson_euclidean | 0.715 |
| spearman_euclidean | 0.7055 |
| pearson_dot | 0.4317 |
| spearman_dot | 0.4293 |
| pearson_max | 0.715 |
| spearman_max | 0.7189 |
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### sentence-transformers/all-nli
* Dataset: [sentence-transformers/all-nli](https://huggingface.co/datasets/sentence-transformers/all-nli) at [cc6c526](https://huggingface.co/datasets/sentence-transformers/all-nli/tree/cc6c526380e29912b5c6fa03682da4daf773c013)
* Size: 10,000 training samples
* Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
* Approximate statistics based on the first 1000 samples:
| | anchor | positive | negative |
|:--------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|
| type | string | string | string |
| details | <ul><li>min: 7 tokens</li><li>mean: 10.38 tokens</li><li>max: 45 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 12.8 tokens</li><li>max: 39 tokens</li></ul> | <ul><li>min: 6 tokens</li><li>mean: 13.4 tokens</li><li>max: 50 tokens</li></ul> |
* Samples:
| anchor | positive | negative |
|:---------------------------------------------------------------------------|:-------------------------------------------------|:-----------------------------------------------------------|
| <code>A person on a horse jumps over a broken down airplane.</code> | <code>A person is outdoors, on a horse.</code> | <code>A person is at a diner, ordering an omelette.</code> |
| <code>Children smiling and waving at camera</code> | <code>There are children present</code> | <code>The kids are frowning</code> |
| <code>A boy is jumping on skateboard in the middle of a red bridge.</code> | <code>The boy does a skateboarding trick.</code> | <code>The boy skates down the sidewalk.</code> |
* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 20.0,
"similarity_fct": "cos_sim"
}
```
### Evaluation Dataset
#### sentence-transformers/all-nli
* Dataset: [sentence-transformers/all-nli](https://huggingface.co/datasets/sentence-transformers/all-nli) at [cc6c526](https://huggingface.co/datasets/sentence-transformers/all-nli/tree/cc6c526380e29912b5c6fa03682da4daf773c013)
* Size: 1,000 evaluation samples
* Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
* Approximate statistics based on the first 1000 samples:
| | anchor | positive | negative |
|:--------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|
| type | string | string | string |
| details | <ul><li>min: 6 tokens</li><li>mean: 18.02 tokens</li><li>max: 66 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 9.81 tokens</li><li>max: 29 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 10.37 tokens</li><li>max: 29 tokens</li></ul> |
* Samples:
| anchor | positive | negative |
|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------|:--------------------------------------------------------|
| <code>Two women are embracing while holding to go packages.</code> | <code>Two woman are holding packages.</code> | <code>The men are fighting outside a deli.</code> |
| <code>Two young children in blue jerseys, one with the number 9 and one with the number 2 are standing on wooden steps in a bathroom and washing their hands in a sink.</code> | <code>Two kids in numbered jerseys wash their hands.</code> | <code>Two kids in jackets walk to school.</code> |
| <code>A man selling donuts to a customer during a world exhibition event held in the city of Angeles</code> | <code>A man selling donuts to a customer.</code> | <code>A woman drinks her coffee in a small cafe.</code> |
* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 20.0,
"similarity_fct": "cos_sim"
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: steps
- `per_device_train_batch_size`: 128
- `per_device_eval_batch_size`: 128
- `num_train_epochs`: 1
- `warmup_ratio`: 0.1
- `fp16`: True
- `batch_sampler`: no_duplicates
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: steps
- `prediction_loss_only`: False
- `per_device_train_batch_size`: 128
- `per_device_eval_batch_size`: 128
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 1
- `eval_accumulation_steps`: None
- `learning_rate`: 5e-05
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1.0
- `num_train_epochs`: 1
- `max_steps`: -1
- `lr_scheduler_type`: linear
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.1
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: False
- `fp16`: True
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: None
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: False
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: None
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: False
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `dispatch_batches`: None
- `split_batches`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_sampler`: no_duplicates
- `multi_dataset_batch_sampler`: proportional
</details>
### Training Logs
| Epoch | Step | loss | sts-dev_spearman_cosine | sts-test_spearman_cosine |
|:------:|:----:|:------:|:-----------------------:|:------------------------:|
| 0 | 0 | - | 0.6375 | - |
| 0.1266 | 10 | 2.9835 | 0.7807 | - |
| 0.2532 | 20 | 1.7048 | 0.7782 | - |
| 0.3797 | 30 | 1.6657 | 0.7847 | - |
| 0.5063 | 40 | 1.7352 | 0.7900 | - |
| 0.6329 | 50 | 1.6400 | 0.7863 | - |
| 0.7595 | 60 | 1.7281 | 0.7820 | - |
| 0.8861 | 70 | 1.7066 | 0.7815 | - |
| 1.0 | 79 | - | - | 0.7189 |
### Environmental Impact
Carbon emissions were measured using [CodeCarbon](https://github.com/mlco2/codecarbon).
- **Energy Consumed**: 0.004 kWh
- **Carbon Emitted**: 0.001 kg of CO2
- **Hours Used**: 0.02 hours
### Training Hardware
- **On Cloud**: No
- **GPU Model**: 1 x NVIDIA GeForce RTX 3090
- **CPU Model**: 13th Gen Intel(R) Core(TM) i7-13700K
- **RAM Size**: 31.78 GB
### Framework Versions
- Python: 3.11.6
- Sentence Transformers: 3.0.0.dev0
- Transformers: 4.41.0.dev0
- PyTorch: 2.3.0+cu121
- Accelerate: 0.26.1
- Datasets: 2.18.0
- Tokenizers: 0.19.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### MultipleNegativesRankingLoss
```bibtex
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION",
"SEMANTIC_SIMILARITY"
] | Non_BioNLP |
akshitha-k/all-MiniLM-L6-v2-stsb | akshitha-k | sentence-similarity | [
"sentence-transformers",
"safetensors",
"bert",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:5749",
"loss:CosineSimilarityLoss",
"arxiv:1908.10084",
"base_model:sentence-transformers/all-MiniLM-L6-v2",
"base_model:finetune:sentence-transformers/all-MiniLM-L6-v2",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,731,273,262,000 | 2024-11-10T21:14:29 | 7 | 0 | ---
base_model: sentence-transformers/all-MiniLM-L6-v2
library_name: sentence-transformers
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:5749
- loss:CosineSimilarityLoss
widget:
- source_sentence: A girl is styling her hair.
sentences:
- China's online population rises to 618 mln
- A girl is filing her nails.
- A woman is slicing a pepper.
- source_sentence: Australian among four on plane missing in Indonesia
sentences:
- Woman dies in Co Cork house fire
- '''No plans'' to resettle Syrian refugees in the UK'
- Iranian painter Mansoureh Hosseini dies
- source_sentence: West hails Syria opposition vote to join peace talks
sentences:
- Asteroid passes Earth in fly-by
- GlaxoSmithKline, the UK drugmaker, has said it would cut off supplies to Canadian
stores shipping drugs to the US.
- Syrian opposition to name delegation for talks
- source_sentence: Obama signs up for Obamacare
sentences:
- Americans scramble to sign up for Obamacare by deadline
- A girl wearing a red blouse riding a brown horse.
- The study also found that skin cancer nearly tripled in Norway and Sweden since
the 1950s.
- source_sentence: A clear plastic chair in front of a bookcase.
sentences:
- A woman with a white horse.
- a clear plastic chair in front of book shelves.
- A herd of caribou are crossing a road.
---
# SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2). It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [sentence-transformers/all-MiniLM-L6-v2](https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2) <!-- at revision fa97f6e7cb1a59073dff9e6b13e2715cf7475ac9 -->
- **Maximum Sequence Length:** 256 tokens
- **Output Dimensionality:** 384 tokens
- **Similarity Function:** Cosine Similarity
<!-- - **Training Dataset:** Unknown -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 256, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("akshitha-k/all-MiniLM-L6-v2-stsb")
# Run inference
sentences = [
'A clear plastic chair in front of a bookcase.',
'a clear plastic chair in front of book shelves.',
'A woman with a white horse.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### Unnamed Dataset
* Size: 5,749 training samples
* Columns: <code>sentence_0</code>, <code>sentence_1</code>, and <code>label</code>
* Approximate statistics based on the first 1000 samples:
| | sentence_0 | sentence_1 | label |
|:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
| type | string | string | float |
| details | <ul><li>min: 6 tokens</li><li>mean: 14.34 tokens</li><li>max: 44 tokens</li></ul> | <ul><li>min: 5 tokens</li><li>mean: 14.31 tokens</li><li>max: 45 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.54</li><li>max: 1.0</li></ul> |
* Samples:
| sentence_0 | sentence_1 | label |
|:------------------------------------------------------------------------------|:--------------------------------------------------------------|:-----------------|
| <code>U.N. rights chief presses Egypt on Mursi detention</code> | <code>UN Rights Chief Presses Egypt on Morsi Detention</code> | <code>1.0</code> |
| <code>Someone is slicing an onion.</code> | <code>Someoen is peeling a potato.</code> | <code>0.2</code> |
| <code>A young boy in a white dress shirt is playing on a grassy plain.</code> | <code>A woman is getting her hair done at a salon.</code> | <code>0.0</code> |
* Loss: [<code>CosineSimilarityLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosinesimilarityloss) with these parameters:
```json
{
"loss_fct": "torch.nn.modules.loss.MSELoss"
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `per_device_train_batch_size`: 16
- `per_device_eval_batch_size`: 16
- `num_train_epochs`: 20
- `multi_dataset_batch_sampler`: round_robin
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: no
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 16
- `per_device_eval_batch_size`: 16
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 1
- `eval_accumulation_steps`: None
- `torch_empty_cache_steps`: None
- `learning_rate`: 5e-05
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1
- `num_train_epochs`: 20
- `max_steps`: -1
- `lr_scheduler_type`: linear
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.0
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: False
- `fp16`: False
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: None
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: False
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: False
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `dispatch_batches`: None
- `split_batches`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `eval_on_start`: False
- `eval_use_gather_object`: False
- `batch_sampler`: batch_sampler
- `multi_dataset_batch_sampler`: round_robin
</details>
### Training Logs
| Epoch | Step | Training Loss |
|:-------:|:----:|:-------------:|
| 1.3889 | 500 | 0.0295 |
| 2.7778 | 1000 | 0.0242 |
| 4.1667 | 1500 | 0.0218 |
| 5.5556 | 2000 | 0.0198 |
| 6.9444 | 2500 | 0.0175 |
| 8.3333 | 3000 | 0.0157 |
| 9.7222 | 3500 | 0.0135 |
| 11.1111 | 4000 | 0.0119 |
| 12.5 | 4500 | 0.0104 |
| 13.8889 | 5000 | 0.0088 |
| 15.2778 | 5500 | 0.0074 |
| 16.6667 | 6000 | 0.0063 |
| 18.0556 | 6500 | 0.0056 |
| 19.4444 | 7000 | 0.0049 |
### Framework Versions
- Python: 3.10.12
- Sentence Transformers: 3.2.1
- Transformers: 4.44.2
- PyTorch: 2.5.0+cu121
- Accelerate: 0.34.2
- Datasets: 3.1.0
- Tokenizers: 0.19.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
Aryanne/Bling-Sheared-Llama-1.3B-0.1-gguf | Aryanne | null | [
"gguf",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | 1,698,170,639,000 | 2023-10-24T18:52:51 | 159 | 4 | ---
license: apache-2.0
---
Some GGUF v2 quantizations of the model [llmware/bling-sheared-llama-1.3b-0.1](https://huggingface.co/llmware/bling-sheared-llama-1.3b-0.1)
bling-sheared-llama-1.3b-0.1 is part of the BLING ("Best Little Instruction-following No-GPU-required") model series, instruct trained on top of a Sheared-LLaMA-1.3B base model.
BLING models are fine-tuned with distilled high-quality custom instruct datasets, targeted at a specific subset of instruct tasks with
the objective of providing a high-quality Instruct model that is 'inference-ready' on a CPU laptop even
without using any advanced quantization optimizations.
### Model Description
- **Developed by:** llmware
- **Model type:** Instruct-trained decoder
- **Language(s) (NLP):** English
- **License:** Apache 2.0
- **Finetuned from model [optional]:** princeton-nlp/Sheared-LLaMA-1.3B
## Uses
The intended use of BLING models is two-fold:
1. Provide high-quality Instruct models that can run on a laptop for local testing. We have found it extremely useful when building a
proof-of-concept, or working with sensitive enterprise data that must be closely guarded, especially in RAG use cases.
2. Push the state of the art for smaller Instruct-following models in the sub-7B parameter range, especially 1B-3B, as single-purpose
automation tools for specific tasks through targeted fine-tuning datasets and focused "instruction" tasks.
## Prompt Format
```
<human>: Anything that you want to say
<bot:
```
or
```
<human>: Context
Instruction/Question
<bot:
```
### Direct Use
BLING is designed for enterprise automation use cases, especially in knowledge-intensive industries, such as financial services,
legal and regulatory industries with complex information sources. Rather than try to be "all things to all people," BLING models try to focus on a narrower set of Instructions more suitable to a ~1B parameter GPT model.
BLING is ideal for rapid prototyping, testing, and the ability to perform an end-to-end workflow locally on a laptop without
having to send sensitive information over an Internet-based API.
The first BLING models have been trained for common RAG scenarios, specifically: question-answering, key-value extraction, and basic summarization as the core instruction types
without the need for a lot of complex instruction verbiage - provide a text passage context, ask questions, and get clear fact-based responses.
## Bias, Risks, and Limitations
Any model can provide inaccurate or incomplete information, and should be used in conjunction with appropriate safeguards and fact-checking mechanisms.
| [
"SUMMARIZATION"
] | Non_BioNLP |
aehrm/redewiedergabe-freeindirect | aehrm | token-classification | [
"flair",
"pytorch",
"token-classification",
"sequence-tagger-model",
"de",
"region:us"
] | 1,684,274,249,000 | 2023-08-23T14:11:55 | 9 | 0 | ---
language: de
tags:
- flair
- token-classification
- sequence-tagger-model
---
# REDEWIEDERGABE Tagger: free indirect STWR
This model is part of an ensemble of binary taggers that recognize German speech, thought and writing representation, that is being used in [LLpro](https://github.com/cophi-wue/LLpro). They can be used to automatically detect and annotate the following 4 types of speech, thought and writing representation in German texts:
| STWR type | Example | Translation |
|--------------------------------|-------------------------------------------------------------------------|----------------------------------------------------------|
| direct | Dann sagte er: **"Ich habe Hunger."** | Then he said: **"I'm hungry."** |
| free indirect ('erlebte Rede', **this tagger**) | Er war ratlos. **Woher sollte er denn hier bloß ein Mittagessen bekommen?** | He was at a loss. **Where should he ever find lunch here?** |
| indirect | Sie fragte, **wo das Essen sei.** | She asked **where the food was.** |
| reported | **Sie sprachen über das Mittagessen.** | **They talked about lunch.** |
The ensemble is trained on the [REDEWIEDERGABE corpus](https://github.com/redewiedergabe/corpus) ([Annotation guidelines](http://redewiedergabe.de/richtlinien/richtlinien.html)), fine-tuning each tagger on the domain-adapted [lkonle/fiction-gbert-large](https://huggingface.co/lkonle/fiction-gbert-large). ([Training Code](https://github.com/cophi-wue/LLpro/blob/main/contrib/train_redewiedergabe.py))
**F1-Scores:**
| STWR type | F1-Score |
|-----------|-----------|
| direct | 90.76 |
| indirect | 79.16 |
| **free indirect (this tagger)** | **58.00** |
| reported | 70.47 |
----
**Demo Usage:**
```python
from flair.data import Sentence
from flair.models import SequenceTagger
sentence = Sentence('Sie sprachen über das Mittagessen. Sie fragte, wo das Essen sei. Woher sollte er das wissen? Dann sagte er: "Ich habe Hunger."')
rwtypes = ['direct', 'indirect', 'freeindirect', 'reported']
for rwtype in rwtypes:
model = SequenceTagger.load(f'aehrm/redewiedergabe-{rwtype}')
model.predict(sentence)
print(rwtype, [ x.data_point.text for x in sentence.get_labels() ])
# >>> direct ['"', 'Ich', 'habe', 'Hunger', '.', '"']
# >>> indirect ['wo', 'das', 'Essen', 'sei', '.']
# >>> freeindirect ['Woher', 'sollte', 'er', 'das', 'wissen', '?']
# >>> reported ['Sie', 'sprachen', 'über', 'das', 'Mittagessen', '.', 'Woher', 'sollte', 'er', 'das', 'wissen', '?']
```
**Cite**:
Please cite the following paper when using this model.
```
@inproceedings{ehrmanntraut-et-al-llpro-2023,
address = {Ingolstadt, Germany},
title = {{LLpro}: A Literary Language Processing Pipeline for {German} Narrative Text},
booktitle = {Proceedings of the 10th Conference on Natural Language Processing ({KONVENS} 2022)},
publisher = {{KONVENS} 2023 Organizers},
author = {Ehrmanntraut, Anton and Konle, Leonard and Jannidis, Fotis},
year = {2023},
}
```
| [
"TRANSLATION"
] | Non_BioNLP |
Helsinki-NLP/opus-mt-sv-umb | Helsinki-NLP | translation | [
"transformers",
"pytorch",
"tf",
"marian",
"text2text-generation",
"translation",
"sv",
"umb",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,646,263,744,000 | 2023-08-16T12:06:25 | 56 | 0 | ---
license: apache-2.0
tags:
- translation
---
### opus-mt-sv-umb
* source languages: sv
* target languages: umb
* OPUS readme: [sv-umb](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/sv-umb/README.md)
* dataset: opus
* model: transformer-align
* pre-processing: normalization + SentencePiece
* download original weights: [opus-2020-01-16.zip](https://object.pouta.csc.fi/OPUS-MT-models/sv-umb/opus-2020-01-16.zip)
* test set translations: [opus-2020-01-16.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/sv-umb/opus-2020-01-16.test.txt)
* test set scores: [opus-2020-01-16.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/sv-umb/opus-2020-01-16.eval.txt)
## Benchmarks
| testset | BLEU | chr-F |
|-----------------------|-------|-------|
| JW300.sv.umb | 20.4 | 0.431 |
| [
"TRANSLATION"
] | Non_BioNLP |
midas/gupshup_e2e_bart | midas | text2text-generation | [
"transformers",
"pytorch",
"bart",
"text2text-generation",
"arxiv:1910.04073",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,646,263,745,000 | 2021-11-14T02:09:24 | 130 | 0 | ---
{}
---
# Gupshup
GupShup: Summarizing Open-Domain Code-Switched Conversations EMNLP 2021
Paper: [https://aclanthology.org/2021.emnlp-main.499.pdf](https://aclanthology.org/2021.emnlp-main.499.pdf)
Github: [https://github.com/midas-research/gupshup](https://github.com/midas-research/gupshup)
### Dataset
Please request for the Gupshup data using [this Google form](https://docs.google.com/forms/d/1zvUk7WcldVF3RCoHdWzQPzPprtSJClrnHoIOYbzaJEI/edit?ts=61381ec0).
Dataset is available for `Hinglish Dilaogues to English Summarization`(h2e) and `English Dialogues to English Summarization`(e2e). For each task, Dialogues/conversastion have `.source`(train.source) as file extension whereas Summary has `.target`(train.target) file extension. ".source" file need to be provided to `input_path` and ".target" file to `reference_path` argument in the scripts.
## Models
All model weights are available on the Huggingface model hub. Users can either directly download these weights in their local and provide this path to `model_name` argument in the scripts or use the provided alias (to `model_name` argument) in scripts directly; this will lead to download weights automatically by scripts.
Model names were aliased in "gupshup_TASK_MODEL" sense, where "TASK" can be h2e,e2e and MODEL can be mbart, pegasus, etc., as listed below.
**1. Hinglish Dialogues to English Summary (h2e)**
| Model | Huggingface Alias |
|---------|-------------------------------------------------------------------------------|
| mBART | [midas/gupshup_h2e_mbart](https://huggingface.co/midas/gupshup_h2e_mbart) |
| PEGASUS | [midas/gupshup_h2e_pegasus](https://huggingface.co/midas/gupshup_h2e_pegasus) |
| T5 MTL | [midas/gupshup_h2e_t5_mtl](https://huggingface.co/midas/gupshup_h2e_t5_mtl) |
| T5 | [midas/gupshup_h2e_t5](https://huggingface.co/midas/gupshup_h2e_t5) |
| BART | [midas/gupshup_h2e_bart](https://huggingface.co/midas/gupshup_h2e_bart) |
| GPT-2 | [midas/gupshup_h2e_gpt](https://huggingface.co/midas/gupshup_h2e_gpt) |
**2. English Dialogues to English Summary (e2e)**
| Model | Huggingface Alias |
|---------|-------------------------------------------------------------------------------|
| mBART | [midas/gupshup_e2e_mbart](https://huggingface.co/midas/gupshup_e2e_mbart) |
| PEGASUS | [midas/gupshup_e2e_pegasus](https://huggingface.co/midas/gupshup_e2e_pegasus) |
| T5 MTL | [midas/gupshup_e2e_t5_mtl](https://huggingface.co/midas/gupshup_e2e_t5_mtl) |
| T5 | [midas/gupshup_e2e_t5](https://huggingface.co/midas/gupshup_e2e_t5) |
| BART | [midas/gupshup_e2e_bart](https://huggingface.co/midas/gupshup_e2e_bart) |
| GPT-2 | [midas/gupshup_e2e_gpt](https://huggingface.co/midas/gupshup_e2e_gpt) |
## Inference
### Using command line
1. Clone this repo and create a python virtual environment (https://docs.python.org/3/library/venv.html). Install the required packages using
```
git clone https://github.com/midas-research/gupshup.git
pip install -r requirements.txt
```
2. run_eval script has the following arguments.
* **model_name** : Path or alias to one of our models available on Huggingface as listed above.
* **input_path** : Source file or path to file containing conversations, which will be summarized.
* **save_path** : File path where to save summaries generated by the model.
* **reference_path** : Target file or path to file containing summaries, used to calculate matrices.
* **score_path** : File path where to save scores.
* **bs** : Batch size
* **device**: Cuda devices to use.
Please make sure you have downloaded the Gupshup dataset using the above google form and provide the correct path to these files in the argument's `input_path` and `refrence_path.` Or you can simply put `test.source` and `test.target` in `data/h2e/`(hinglish to english) or `data/e2e/`(english to english) folder. For example, to generate English summaries from Hinglish dialogues using the mbart model, run the following command
```
python run_eval.py \
--model_name midas/gupshup_h2e_mbart \
--input_path data/h2e/test.source \
--save_path generated_summary.txt \
--reference_path data/h2e/test.target \
--score_path scores.txt \
--bs 8
```
Another example, to generate English summaries from English dialogues using the Pegasus model
```
python run_eval.py \
--model_name midas/gupshup_e2e_pegasus \
--input_path data/e2e/test.source \
--save_path generated_summary.txt \
--reference_path data/e2e/test.target \
--score_path scores.txt \
--bs 8
```
Please create an issue if you are facing any difficulties in replicating the results.
### References
Please cite [[1]](https://arxiv.org/abs/1910.04073) if you found the resources in this repository useful.
[1] Mehnaz, Laiba, Debanjan Mahata, Rakesh Gosangi, Uma Sushmitha Gunturi, Riya Jain, Gauri Gupta, Amardeep Kumar, Isabelle G. Lee, Anish Acharya, and Rajiv Shah. [*GupShup: Summarizing Open-Domain Code-Switched Conversations*](https://aclanthology.org/2021.emnlp-main.499.pdf)
```
@inproceedings{mehnaz2021gupshup,
title={GupShup: Summarizing Open-Domain Code-Switched Conversations},
author={Mehnaz, Laiba and Mahata, Debanjan and Gosangi, Rakesh and Gunturi, Uma Sushmitha and Jain, Riya and Gupta, Gauri and Kumar, Amardeep and Lee, Isabelle G and Acharya, Anish and Shah, Rajiv},
booktitle={Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing},
pages={6177--6192},
year={2021}
}
```
| [
"SUMMARIZATION"
] | Non_BioNLP |
TransferGraph/SetFit_distilbert-base-uncased__sst2__train-16-0-finetuned-lora-tweet_eval_irony | TransferGraph | text-classification | [
"peft",
"safetensors",
"parquet",
"text-classification",
"dataset:tweet_eval",
"base_model:SetFit/distilbert-base-uncased__sst2__train-16-0",
"base_model:adapter:SetFit/distilbert-base-uncased__sst2__train-16-0",
"license:apache-2.0",
"model-index",
"region:us"
] | 1,709,053,281,000 | 2024-02-27T17:01:26 | 0 | 0 | ---
base_model: SetFit/distilbert-base-uncased__sst2__train-16-0
datasets:
- tweet_eval
library_name: peft
license: apache-2.0
metrics:
- accuracy
tags:
- parquet
- text-classification
model-index:
- name: SetFit_distilbert-base-uncased__sst2__train-16-0-finetuned-lora-tweet_eval_irony
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: tweet_eval
type: tweet_eval
config: irony
split: validation
args: irony
metrics:
- type: accuracy
value: 0.6858638743455497
name: accuracy
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# SetFit_distilbert-base-uncased__sst2__train-16-0-finetuned-lora-tweet_eval_irony
This model is a fine-tuned version of [SetFit/distilbert-base-uncased__sst2__train-16-0](https://huggingface.co/SetFit/distilbert-base-uncased__sst2__train-16-0) on the tweet_eval dataset.
It achieves the following results on the evaluation set:
- accuracy: 0.6859
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0005
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| accuracy | train_loss | epoch |
|:--------:|:----------:|:-----:|
| 0.5215 | None | 0 |
| 0.5937 | 0.6758 | 0 |
| 0.6115 | 0.6295 | 1 |
| 0.6555 | 0.5843 | 2 |
| 0.6838 | 0.5587 | 3 |
| 0.6712 | 0.5388 | 4 |
| 0.6534 | 0.5176 | 5 |
| 0.6785 | 0.5086 | 6 |
| 0.6859 | 0.5013 | 7 |
### Framework versions
- PEFT 0.8.2
- Transformers 4.37.2
- Pytorch 2.2.0
- Datasets 2.16.1
- Tokenizers 0.15.2 | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
tartuNLP/llammas-prelim | tartuNLP | text-generation | [
"transformers",
"pytorch",
"safetensors",
"llama",
"text-generation",
"et",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,699,017,294,000 | 2023-11-14T12:17:40 | 9 | 1 | ---
language:
- et
widget:
- text: 'Mida sa tead Juhan Liivi kohta? Vastus:'
---
Llama-2-7B finetuned in three stages:
1. 1B tokens of CulturaX (75% Estonain, 25% English)
2. 1M English->Estonian sentence-pairs from CCMatrix (500000), WikiMatrix (400000), Europarl (50000), and OpenSubtitles (50000) as Alpaca-style translation instructions
3. Alpaca-cleaned and Alpaca-est (both ~50000 instructions)
Alpaca-est is an instruction dataset generated for Estonian with *gpt-3.5-turbo-0613*, following Alpaca. | [
"TRANSLATION"
] | Non_BioNLP |
rambodazimi/distilroberta-base-finetuned-LoRA-MRPC | rambodazimi | null | [
"safetensors",
"generated_from_trainer",
"dataset:glue",
"license:apache-2.0",
"model-index",
"region:us"
] | 1,725,642,888,000 | 2024-09-06T17:16:42 | 0 | 0 | ---
datasets:
- glue
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: distilroberta-base-finetuned-LoRA-MRPC
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: glue
type: glue
args: mrpc
metrics:
- type: accuracy
value: 0.8455882352941176
name: Accuracy
- type: f1
value: 0.8911917098445595
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilroberta-base-finetuned-lora-mrpc
This model is a fine-tuned version of [distilroberta-base](https://huggingface.co/distilbert/distilroberta-base) on the glue dataset.
It achieves the following results on the evaluation set:
- Accuracy: 0.8456
- F1: 0.8912
- trainable model parameters: 887042
- all model parameters: 83006980
- percentage of trainable model parameters: 1.07%
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-04
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- weight_decay: 0.01
- rank: 16
- lora_alpha: 32
- lora_dropout: 0.05
- num_epochs: 4 | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
marefa-nlp/marefa-ner | marefa-nlp | token-classification | [
"transformers",
"pytorch",
"xlm-roberta",
"token-classification",
"ar",
"dataset:Marefa-NER",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,646,263,745,000 | 2021-12-04T05:21:57 | 3,785 | 23 | ---
datasets:
- Marefa-NER
language: ar
widget:
- text: في استاد القاهرة، بدأ حفل افتتاح بطولة كأس الأمم الأفريقية بحضور رئيس الجمهورية
و رئيس الاتحاد الدولي لكرة القدم
---
# Tebyan تبيـان
## Marefa Arabic Named Entity Recognition Model
## نموذج المعرفة لتصنيف أجزاء النص
<p align="center">
<img src="https://huggingface.co/marefa-nlp/marefa-ner/resolve/main/assets/marefa-tebyan-banner.png" alt="Marfa Arabic NER Model" width="600"/>
</p?
---------
**Version**: 1.3
**Last Update:** 3-12-2021
## Model description
**Marefa-NER** is a Large Arabic Named Entity Recognition (NER) model built on a completely new dataset and targets to extract up to 9 different types of entities
```
Person, Location, Organization, Nationality, Job, Product, Event, Time, Art-Work
```
نموذج المعرفة لتصنيف أجزاء النص. نموذج جديد كليا من حيث البيانات المستخدمة في تدريب النموذج.
كذلك يستهدف النموذج تصنيف حتى 9 أنواع مختلفة من أجزاء النص
```
شخص - مكان - منظمة - جنسية - وظيفة - منتج - حدث - توقيت - عمل إبداعي
```
## How to use كيف تستخدم النموذج
*You can test the model quickly by checking this [Colab notebook](https://colab.research.google.com/drive/1OGp9Wgm-oBM5BBhTLx6Qow4dNRSJZ-F5?usp=sharing)*
----
Install the following Python packages
`$ pip3 install transformers==4.8.0 nltk==3.5 protobuf==3.15.3 torch==1.9.0 `
> If you are using `Google Colab`, please restart your runtime after installing the packages.
-----------
```python
from transformers import AutoTokenizer, AutoModelForTokenClassification
import torch
import numpy as np
import nltk
nltk.download('punkt')
from nltk.tokenize import word_tokenize
custom_labels = ["O", "B-job", "I-job", "B-nationality", "B-person", "I-person", "B-location","B-time", "I-time", "B-event", "I-event", "B-organization", "I-organization", "I-location", "I-nationality", "B-product", "I-product", "B-artwork", "I-artwork"]
def _extract_ner(text: str, model: AutoModelForTokenClassification,
tokenizer: AutoTokenizer, start_token: str="▁"):
tokenized_sentence = tokenizer([text], padding=True, truncation=True, return_tensors="pt")
tokenized_sentences = tokenized_sentence['input_ids'].numpy()
with torch.no_grad():
output = model(**tokenized_sentence)
last_hidden_states = output[0].numpy()
label_indices = np.argmax(last_hidden_states[0], axis=1)
tokens = tokenizer.convert_ids_to_tokens(tokenized_sentences[0])
special_tags = set(tokenizer.special_tokens_map.values())
grouped_tokens = []
for token, label_idx in zip(tokens, label_indices):
if token not in special_tags:
if not token.startswith(start_token) and len(token.replace(start_token,"").strip()) > 0:
grouped_tokens[-1]["token"] += token
else:
grouped_tokens.append({"token": token, "label": custom_labels[label_idx]})
# extract entities
ents = []
prev_label = "O"
for token in grouped_tokens:
label = token["label"].replace("I-","").replace("B-","")
if token["label"] != "O":
if label != prev_label:
ents.append({"token": [token["token"]], "label": label})
else:
ents[-1]["token"].append(token["token"])
prev_label = label
# group tokens
ents = [{"token": "".join(rec["token"]).replace(start_token," ").strip(), "label": rec["label"]} for rec in ents ]
return ents
model_cp = "marefa-nlp/marefa-ner"
tokenizer = AutoTokenizer.from_pretrained(model_cp)
model = AutoModelForTokenClassification.from_pretrained(model_cp, num_labels=len(custom_labels))
samples = [
"تلقى تعليمه في الكتاب ثم انضم الى الأزهر عام 1873م. تعلم على يد السيد جمال الدين الأفغاني والشيخ محمد عبده",
"بعد عودته إلى القاهرة، التحق نجيب الريحاني فرقة جورج أبيض، الذي كان قد ضمَّ - قُبيل ذلك - فرقته إلى فرقة سلامة حجازي . و منها ذاع صيته",
"في استاد القاهرة، قام حفل افتتاح بطولة كأس الأمم الأفريقية بحضور رئيس الجمهورية و رئيس الاتحاد الدولي لكرة القدم",
"من فضلك أرسل هذا البريد الى صديقي جلال الدين في تمام الساعة الخامسة صباحا في يوم الثلاثاء القادم",
"امبارح اتفرجت على مباراة مانشستر يونايتد مع ريال مدريد في غياب الدون كرستيانو رونالدو",
"لا تنسى تصحيني الساعة سبعة, و ضيف في الجدول اني احضر مباراة نادي النصر غدا",
]
# [optional]
samples = [ " ".join(word_tokenize(sample.strip())) for sample in samples if sample.strip() != "" ]
for sample in samples:
ents = _extract_ner(text=sample, model=model, tokenizer=tokenizer, start_token="▁")
print(sample)
for ent in ents:
print("\t",ent["token"],"==>",ent["label"])
print("========\n")
```
Output
```
تلقى تعليمه في الكتاب ثم انضم الى الأزهر عام 1873م . تعلم على يد السيد جمال الدين الأفغاني والشيخ محمد عبده
الأزهر ==> organization
عام 1873م ==> time
السيد جمال الدين الأفغاني ==> person
محمد عبده ==> person
========
بعد عودته إلى القاهرة، التحق نجيب الريحاني فرقة جورج أبيض، الذي كان قد ضمَّ - قُبيل ذلك - فرقته إلى فرقة سلامة حجازي . و منها ذاع صيته
القاهرة، ==> location
نجيب الريحاني ==> person
فرقة جورج أبيض، ==> organization
فرقة سلامة حجازي ==> organization
========
في استاد القاهرة، قام حفل افتتاح بطولة كأس الأمم الأفريقية بحضور رئيس الجمهورية و رئيس الاتحاد الدولي لكرة القدم
استاد القاهرة، ==> location
بطولة كأس الأمم الأفريقية ==> event
رئيس الجمهورية ==> job
رئيس ==> job
الاتحاد الدولي لكرة القدم ==> organization
========
من فضلك أرسل هذا البريد الى صديقي جلال الدين في تمام الساعة الخامسة صباحا في يوم الثلاثاء القادم
جلال الدين ==> person
الساعة الخامسة صباحا ==> time
يوم الثلاثاء القادم ==> time
========
امبارح اتفرجت على مباراة مانشستر يونايتد مع ريال مدريد في غياب الدون كرستيانو رونالدو
مانشستر يونايتد ==> organization
ريال مدريد ==> organization
كرستيانو رونالدو ==> person
========
لا تنسى تصحيني الساعة سبعة , و ضيف في الجدول اني احضر مباراة نادي النصر غدا
الساعة سبعة ==> time
نادي النصر ==> organization
غدا ==> time
========
```
## Fine-Tuning
Check this [notebook](https://colab.research.google.com/drive/1WUYrnmDFFEItqGMvbyjqZEJJqwU7xQR-?usp=sharing) to fine-tune the NER model
## Evaluation
We tested the model agains a test set of 1959 sentences. The results is in the follwing table
| type | f1-score | precision | recall | support |
|:-------------|-----------:|------------:|---------:|----------:|
| person | 0.93298 | 0.931479 | 0.934487 | 4335 |
| location | 0.891537 | 0.896926 | 0.886212 | 4939 |
| time | 0.873003 | 0.876087 | 0.869941 | 1853 |
| nationality | 0.871246 | 0.843153 | 0.901277 | 2350 |
| job | 0.837656 | 0.79912 | 0.880097 | 2477 |
| organization | 0.781317 | 0.773328 | 0.789474 | 2299 |
| event | 0.686695 | 0.733945 | 0.645161 | 744 |
| artwork | 0.653552 | 0.678005 | 0.630802 | 474 |
| product | 0.625483 | 0.553531 | 0.718935 | 338 |
| **weighted avg** | 0.859008 | 0.852365 | 0.86703 | 19809 |
| **micro avg** | 0.858771 | 0.850669 | 0.86703 | 19809 |
| **macro avg** | 0.79483 | 0.787286 | 0.806265 | 19809 |
## Acknowledgment شكر و تقدير
قام بإعداد البيانات التي تم تدريب النموذج عليها, مجموعة من المتطوعين الذين قضوا ساعات يقومون بتنقيح البيانات و مراجعتها
- على سيد عبد الحفيظ - إشراف
- نرمين محمد عطيه
- صلاح خيرالله
- احمد علي عبدربه
- عمر بن عبد العزيز سليمان
- محمد ابراهيم الجمال
- عبدالرحمن سلامه خلف
- إبراهيم كمال محمد سليمان
- حسن مصطفى حسن
- أحمد فتحي سيد
- عثمان مندو
- عارف الشريف
- أميرة محمد محمود
- حسن سعيد حسن
- عبد العزيز علي البغدادي
- واثق عبدالملك الشويطر
- عمرو رمضان عقل الحفناوي
- حسام الدين أحمد على
- أسامه أحمد محمد محمد
- حاتم محمد المفتي
- عبد الله دردير
- أدهم البغدادي
- أحمد صبري
- عبدالوهاب محمد محمد
- أحمد محمد عوض | [
"NAMED_ENTITY_RECOGNITION"
] | Non_BioNLP |
learn3r/longt5_xl_sfd_bp_20 | learn3r | text2text-generation | [
"transformers",
"pytorch",
"longt5",
"text2text-generation",
"generated_from_trainer",
"dataset:learn3r/summ_screen_fd_bp",
"base_model:google/long-t5-tglobal-xl",
"base_model:finetune:google/long-t5-tglobal-xl",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,698,974,157,000 | 2023-11-04T06:54:50 | 18 | 0 | ---
base_model: google/long-t5-tglobal-xl
datasets:
- learn3r/summ_screen_fd_bp
license: apache-2.0
metrics:
- rouge
tags:
- generated_from_trainer
model-index:
- name: longt5_xl_sfd_bp_20
results:
- task:
type: summarization
name: Summarization
dataset:
name: learn3r/summ_screen_fd_bp
type: learn3r/summ_screen_fd_bp
metrics:
- type: rouge
value: 22.11
name: Rouge1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# longt5_xl_sfd_bp_20
This model is a fine-tuned version of [google/long-t5-tglobal-xl](https://huggingface.co/google/long-t5-tglobal-xl) on the learn3r/summ_screen_fd_bp dataset.
It achieves the following results on the evaluation set:
- Loss: 1.5032
- Rouge1: 22.11
- Rouge2: 7.544
- Rougel: 19.7035
- Rougelsum: 20.2813
- Gen Len: 497.8783
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.001
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 32
- total_train_batch_size: 256
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 20.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:|
| 2.3973 | 0.97 | 14 | 1.9074 | 10.6164 | 2.4585 | 10.4856 | 9.8193 | 511.0 |
| 1.9188 | 1.95 | 28 | 1.7082 | 17.4258 | 4.2128 | 16.5213 | 15.8377 | 511.0 |
| 1.4297 | 2.99 | 43 | 1.5073 | 18.6504 | 5.4242 | 17.2648 | 17.0203 | 506.7745 |
| 1.2759 | 3.97 | 57 | 1.5032 | 22.11 | 7.544 | 19.7035 | 20.2813 | 497.8783 |
| 1.1421 | 4.94 | 71 | 1.5462 | 20.6049 | 6.7146 | 18.5084 | 19.0876 | 503.6024 |
| 0.9605 | 5.98 | 86 | 1.6233 | 22.6777 | 7.9362 | 18.7936 | 21.41 | 510.2730 |
| 0.8082 | 6.96 | 100 | 1.7575 | 26.5338 | 9.9474 | 20.3789 | 25.0767 | 511.0 |
| 0.664 | 8.0 | 115 | 1.7702 | 35.1918 | 13.7223 | 26.1763 | 33.3997 | 329.7151 |
| 0.5471 | 8.97 | 129 | 1.9383 | 27.0414 | 10.4166 | 20.1803 | 25.6283 | 506.8279 |
| 0.4349 | 9.95 | 143 | 1.9608 | 29.5613 | 11.7633 | 22.7176 | 27.9563 | 454.7033 |
| 0.4338 | 10.99 | 158 | 2.1197 | 31.2004 | 12.8569 | 22.1282 | 29.8827 | 493.3234 |
| 0.2887 | 11.97 | 172 | 2.1205 | 34.9566 | 13.8574 | 25.1764 | 33.2914 | 381.3591 |
| 0.2753 | 12.94 | 186 | 2.4299 | 36.3877 | 13.8584 | 25.7829 | 34.8601 | 338.7240 |
| 0.2114 | 13.98 | 201 | 2.5799 | 39.7535 | 16.1209 | 27.8512 | 37.8553 | 302.4837 |
| 0.1805 | 14.96 | 215 | 2.6123 | 33.3254 | 13.0868 | 23.3214 | 31.7901 | 442.9258 |
| 0.1543 | 16.0 | 230 | 2.5635 | 31.7816 | 13.1085 | 22.9117 | 30.2286 | 463.0801 |
| 0.5166 | 16.97 | 244 | 2.5134 | 30.3969 | 12.1295 | 21.6616 | 28.7606 | 511.0 |
| 0.1117 | 17.95 | 258 | 2.8109 | 35.336 | 14.9492 | 24.1938 | 33.822 | 431.1157 |
| 0.0895 | 18.99 | 273 | 2.7577 | 41.0982 | 16.3935 | 28.1073 | 39.1641 | 240.1365 |
| 0.0779 | 19.48 | 280 | 2.8927 | 32.7788 | 13.9352 | 22.5175 | 31.548 | 488.5134 |
### Framework versions
- Transformers 4.34.1
- Pytorch 2.1.0+cu121
- Datasets 2.14.5
- Tokenizers 0.14.1
| [
"SUMMARIZATION"
] | Non_BioNLP |
Yanis23/sparql-translation | Yanis23 | translation | [
"transformers",
"pytorch",
"tensorboard",
"bart",
"text2text-generation",
"translation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,682,343,918,000 | 2023-04-24T19:09:42 | 28 | 0 | ---
license: apache-2.0
tags:
- translation
- generated_from_trainer
model-index:
- name: sparql-translation
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# sparql-translation
This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on the None dataset.
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
| [
"TRANSLATION"
] | Non_BioNLP |
Hoax0930/marian-finetuned-kftt_kde4-en-to-ja | Hoax0930 | translation | [
"transformers",
"pytorch",
"tensorboard",
"marian",
"text2text-generation",
"translation",
"generated_from_trainer",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,664,867,663,000 | 2022-10-04T08:25:10 | 98 | 0 | ---
license: apache-2.0
metrics:
- bleu
tags:
- translation
- generated_from_trainer
model-index:
- name: marian-finetuned-kftt_kde4-en-to-ja
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# marian-finetuned-kftt_kde4-en-to-ja
This model is a fine-tuned version of [Hoax0930/kyoto_marian_mod_2_2_1](https://huggingface.co/Hoax0930/kyoto_marian_mod_2_2_1) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 8.3622
- Bleu: 2.6910
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.22.2
- Pytorch 1.12.1+cu113
- Datasets 2.5.1
- Tokenizers 0.12.1
| [
"TRANSLATION"
] | Non_BioNLP |
gokuls/hBERTv2_new_pretrain_w_init_48_ver2_stsb | gokuls | text-classification | [
"transformers",
"pytorch",
"hybridbert",
"text-classification",
"generated_from_trainer",
"en",
"dataset:glue",
"base_model:gokuls/bert_12_layer_model_v2_complete_training_new_wt_init_48",
"base_model:finetune:gokuls/bert_12_layer_model_v2_complete_training_new_wt_init_48",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,697,611,370,000 | 2023-10-18T06:52:03 | 36 | 0 | ---
base_model: gokuls/bert_12_layer_model_v2_complete_training_new_wt_init_48
datasets:
- glue
language:
- en
metrics:
- spearmanr
tags:
- generated_from_trainer
model-index:
- name: hBERTv2_new_pretrain_w_init_48_ver2_stsb
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: GLUE STSB
type: glue
config: stsb
split: validation
args: stsb
metrics:
- type: spearmanr
value: 0.19761262239980293
name: Spearmanr
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# hBERTv2_new_pretrain_w_init_48_ver2_stsb
This model is a fine-tuned version of [gokuls/bert_12_layer_model_v2_complete_training_new_wt_init_48](https://huggingface.co/gokuls/bert_12_layer_model_v2_complete_training_new_wt_init_48) on the GLUE STSB dataset.
It achieves the following results on the evaluation set:
- Loss: 2.2194
- Pearson: 0.2187
- Spearmanr: 0.1976
- Combined Score: 0.2081
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 4e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 10
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
### Training results
| Training Loss | Epoch | Step | Validation Loss | Pearson | Spearmanr | Combined Score |
|:-------------:|:-----:|:----:|:---------------:|:-------:|:---------:|:--------------:|
| 2.3584 | 1.0 | 90 | 2.3085 | 0.1702 | 0.1471 | 0.1586 |
| 2.0513 | 2.0 | 180 | 2.4060 | 0.1479 | 0.1342 | 0.1411 |
| 1.9851 | 3.0 | 270 | 2.4888 | 0.0897 | 0.1163 | 0.1030 |
| 1.8287 | 4.0 | 360 | 2.7571 | 0.1643 | 0.1827 | 0.1735 |
| 1.6845 | 5.0 | 450 | 2.2194 | 0.2187 | 0.1976 | 0.2081 |
| 1.6892 | 6.0 | 540 | 2.4431 | 0.1882 | 0.1858 | 0.1870 |
| 1.5272 | 7.0 | 630 | 2.6124 | 0.1433 | 0.1572 | 0.1503 |
| 1.402 | 8.0 | 720 | 2.8100 | 0.1605 | 0.1671 | 0.1638 |
| 1.3122 | 9.0 | 810 | 2.7081 | 0.1298 | 0.1428 | 0.1363 |
| 1.187 | 10.0 | 900 | 2.8638 | 0.1724 | 0.1825 | 0.1775 |
### Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
rootacess/distilbert-base-uncased-distilled-clinc | rootacess | text-classification | [
"transformers",
"pytorch",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:clinc_oos",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,679,035,202,000 | 2023-03-17T06:51:16 | 29 | 0 | ---
datasets:
- clinc_oos
license: apache-2.0
metrics:
- accuracy
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-distilled-clinc
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: clinc_oos
type: clinc_oos
config: plus
split: validation
args: plus
metrics:
- type: accuracy
value: 0.937741935483871
name: Accuracy
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-distilled-clinc
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the clinc_oos dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1263
- Accuracy: 0.9377
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 48
- eval_batch_size: 48
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 8
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| No log | 1.0 | 318 | 0.7135 | 0.7110 |
| 0.9811 | 2.0 | 636 | 0.3228 | 0.8561 |
| 0.9811 | 3.0 | 954 | 0.1909 | 0.9094 |
| 0.3187 | 4.0 | 1272 | 0.1517 | 0.9261 |
| 0.1735 | 5.0 | 1590 | 0.1379 | 0.9310 |
| 0.1735 | 6.0 | 1908 | 0.1308 | 0.9342 |
| 0.1414 | 7.0 | 2226 | 0.1275 | 0.9368 |
| 0.1306 | 8.0 | 2544 | 0.1263 | 0.9377 |
### Framework versions
- Transformers 4.27.1
- Pytorch 1.13.1+cu116
- Datasets 2.10.1
- Tokenizers 0.13.2
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
research-dump/bge-base-en-v1.5_wikipedia_r_masked_wikipedia_r_masked | research-dump | text-classification | [
"setfit",
"safetensors",
"bert",
"sentence-transformers",
"text-classification",
"generated_from_setfit_trainer",
"arxiv:2209.11055",
"base_model:BAAI/bge-base-en-v1.5",
"base_model:finetune:BAAI/bge-base-en-v1.5",
"region:us"
] | 1,738,904,700,000 | 2025-02-07T05:05:17 | 9 | 0 | ---
base_model: BAAI/bge-base-en-v1.5
library_name: setfit
metrics:
- accuracy
pipeline_tag: text-classification
tags:
- setfit
- sentence-transformers
- text-classification
- generated_from_setfit_trainer
widget:
- text: 'St. Timothy High School (Cochrane): Luna <3 (She/Her) ( talk ) 04:19, 15
July 2023 (UTC) This discussion has been included in the list of Canada-related
deletion discussions . Luna <3 (She/Her) ( talk ) 04:20, 15 July 2023 (UTC) This
discussion has been included in the deletion sorting lists for the following topics:
Education , Schools , and Christianity . Spiderone (Talk to Spider) 11:10, 15
July 2023 (UTC) High schools are usually the subject of enough verifiable and
reliable published sources to pass WP:GNG . After a quick search I found coverage
in the Calgary Herald and another article about their new principal and added
them to the article. I suspect there are more sources out there to develop the
article. –– Formal Dude (talk) 15:41, 15 July 2023 (UTC) [ reply ] per nom. I
haven''t been able to find anything that would satisfy WP:NSCHOOL . The notion
that high schools are generally notable was discarded quite a while back. Clarityfiend
( talk ) 22:33, 15 July 2023 (UTC) [ reply ] to Calgary Catholic School District
. I have spent some time searching for sources, but nothing significant at this
stage. It could just be WP:TOOSOON - the school is quite new, although it is heading
for its 20th anniversary. All the same, it is not very large, and doesn''t appear
to be independently notable. Per Clarityfiend, secondary schools are no longer
presumed notable, and that is a consensus view from an RfC. However, neither did
FormalDude say that they were - merely that such schools are usually the subject
of sufficient coverage in sources to pass GNG. This is true, but at this point
I do not see that for this case. The Calgary Herald articles have limited distribution,
and an article about appointment of staff at what is essentially a local school
is not sufficient in itself to establish notability. I do feel, however, that
a is a suitable alternative to deletion in this case. Redirects are WP:CHEAP ,
and searching on the name of the school with the place (per the title) is a plausible
search term - particularly by people in the locality. Information on this page
is almost entirely on the target page, and that would be a suitable landing page
for anyone conducting the search. I do not see that deletion prior to is required
- thus page history would be preserved should more sources come to light in the
future, such that this could be expanded into an encylopaedic artice. As it may
just be TOOSOON, it is entirely possible that such an article could be written
one day. Sirfurboy🏄 ( talk ) 14:03, 16 July 2023 (UTC) [ reply ] per Sirfurboy.
IMO this is a better way of organizing schools coverage in general, if only because
it affords a much less inviting target for the various kinds of abuse that school
articles are traditionally subjected to. At any rate it seems like a substantially
better approach here, where there really doesn''t seem to be sourcing currently
available that could support more than a permastub. It can always be spun back
out later if sufficient source material is located. -- Visviva ( talk ) 03:06,
17 July 2023 (UTC) Not going to ! vote for obvious COI reasons, but I would support
a per Sirfurboy and Visviva . Luna <3 (She/Her) ( talk ) 05:15, 19 July 2023 (UTC)
[ reply ] The above discussion is preserved as an archive of the debate. Please
do not modify it. Subsequent comments should be made on the appropriate discussion
page (such as the article''s talk page or in a deletion review ). No further edits
should be made to this page.'
- text: 'Erum Ali : Has styled her husband in some movies and was the lead costume
designer in one movie. All her notability is inherited from her husband and fails
WP:GNG Jupitus Smart 15:15, 21 April 2023 (UTC) This discussion has been included
in the deletion sorting lists for the following topics: Actors and filmmakers
, Fashion , India , and New Zealand . Jupitus Smart 15:15, 21 April 2023 (UTC)
[ reply ] to Abbas (actor) . Not too concerned whether or not she got her start
in films due to nepotism; that''s hallowed tradition in film-industries worldwide
:). But I agree with the nominator that the subject fails WP:GNG at present. The
two main sources cited are an interview with the subject and a short, soft-focused
piece of the type that is typically written and distributed by publicists to accompany
a film opening. A web search didn''t find anything more substantial that would
be needed to establish notability under WP:GNG or WP:CREATIVE . Since the subject
and her career as a fashion designer are already mentioned in Abbas (actor)#Personal
life , a would be justified and should suffice for now. Abecedare ( talk ) 15:59,
21 April 2023 (UTC) This discussion has been included in the deletion sorting
lists for the following topics: Women and Tamil Nadu . Spiderone (Talk to Spider)
17:22, 21 April 2023 (UTC) 10, 22 April 2023 (UTC) [ reply ] Fails GNG but a to
her husband is appropriate as her work is mentioned there. Schwede 66 19:44, 22
April 2023 (UTC) [ reply ] to Abbas (actor) . BLP, fails GNG and BIO. Notability
is not inherited. // Timothy :: talk 01:35, 27 April 2023 (UTC) [ reply ] The
above discussion is preserved as an archive of the debate. Please do not modify
it. Subsequent comments should be made on the appropriate discussion page (such
as the article''s talk page or in a deletion review ). No further edits should
be made to this page.'
- text: 'Simon Sues : – Laundry Pizza 03 ( d c̄ ) 15:37, 12 April 2023 (UTC) This
discussion has been included in the deletion sorting lists for the following topics:
Science fiction and fantasy , Comics and animation , and Anime and manga . – Laundry
Pizza 03 ( d c̄ ) 15:37, 12 April 2023 (UTC) [ reply ] issues not addressed for
8 years; Google only throws up listings and associated official and/or fan social
media accounts, meaning notable neutral sources are unlikely to be found, at most
to Tokyopop as seemingly only notable attached page. BoomboxTestarossa ( talk
) 16:15, 12 April 2023 (UTC) [ reply ] . It''s ok to mass prod such unreferenced
fancruft. -- Piotr Konieczny aka Prokonsul Piotrus | reply here 16:26, 12 April
2023 (UTC) [ reply ] Mass prod was my mistake, I forgot no-one could see that
I''d checked to see if anything was salvageable beforehand and put through too
many at once. I have been put gently right by some kind people, and hopefully
will not have to do anything with that many pages at once again! =) BoomboxTestarossa
( talk ) 20:46, 12 April 2023 (UTC) [ reply ] . ZERO sources used, I''m not wading
through every Google mention with a person named Simon that sues people. TNT this
article. Oaktree b ( talk ) 20:00, 12 April 2023 (UTC) [ reply ] Strongly agree
with what’s been said above. This is a clear . Go4thProsper ( talk ) 08:14, 15
April 2023 (UTC) [ reply ] - literally unsourced and difficult to find anything.
Bearian ( talk ) 19:11, 18 April 2023 (UTC) GNG . Pharaoh of the Wizards ( talk
) 08:25, 19 April 2023 (UTC) [ reply ] The above discussion is preserved as an
archive of the debate. Please do not modify it. Subsequent comments should be
made on the appropriate discussion page (such as the article''s talk page or in
a deletion review ). No further edits should be made to this page.'
- text: 'Selfie Type: Never launched. See WP:CRYSTALBALL . Delta space 42 ( talk •
contribs ) 11:25, 21 December 2023 (UTC) This discussion has been included in
the deletion sorting lists for the following topics: Technology and Software .
Delta space 42 ( talk • contribs ) 11:25, 21 December 2023 (UTC) coverage by BBC
and others establishes notability. Whether it was launched or not is irrelevant.
We have many articles for products and projects that never came to fruition, and
there''s no policy to remove such articles. WP:CRYSTALBALL doesn''t apply, as
this isn''t a prediction, it''s a description of a past event or idea, which still
exists as a notable idea. Owen× ☎ 15:52, 21 December 2023 (UTC) [ reply ] into
Virtual keyboard - changed per discussion below with nom. More than enough sources
to support this as a section in the target. Owen× ☎ 18:06, 21 December 2023 (UTC)
CRYSTALBALL , I meant that it is very unlikely that there will be more info about
this particular product from Samsung in the near future, I just mentioned crystal
ball because it is a nice idea to have "an invisible keyboard", but since this
particular product didn''t launch, we can''t predict that there will be another
product implementing this idea, thus I think it shouldn''t have a standalone article.
Delta space 42 ( talk • contribs ) 17:12, 21 December 2023 (UTC) [ reply ] Also,
coverage by BBC and others establishes notability . Since the only coverage is
about announcement, I believe it is not enough for notability: a burst of coverage
(often around product announcements) does not automatically make a product notable
per WP:NSOFT . Delta space 42 ( talk • contribs ) 17:15, 21 December 2023 (UTC)
[ reply ] I understand your point, and agree that such a burst of coverage does
not automatically establish notability, and you are correct that we are unlikely
to get additional sources about this concept. My claim is that when you take BBC
along with all the other sources, this concept of a product marginally passes
our usual threshold of notability. That said, in an effort to drive to a consensus
here, how would you feel about merging this article to Virtual keyboard ? We certainly
have more than enough sources here for a section in the target page. Owen× ☎ 17:40,
21 December 2023 (UTC) [ reply ] @ OwenX , I think it''s a good idea. I noticed
that there is a section "Optical virtual keyboard" which is exactly what Selfie
Type was about. Delta space 42 ( talk • contribs ) 17:44, 21 December 2023 (UTC)
[ reply ] Excellent. I changed my ! vote above. If you revise the nomination,
we can speedy close this AfD as withdrawn and carry out the as agreed. Owen× ☎
18:06, 21 December 2023 (UTC) [ reply ] The above discussion is preserved as an
archive of the debate. Please do not modify it. Subsequent comments should be
made on the appropriate discussion page (such as the article''s talk page or in
a deletion review ). No further edits should be made to this page.'
- text: 'Viverse: Taking it here to see if there is consensus to it a . – Joe ( talk
) 03:40, 5 July 2023 (UTC) This discussion has been included in the list of Software-related
deletion discussions . – Joe ( talk ) 03:40, 5 July 2023 (UTC) [ reply ] There
has been quite a bit of in-depth coverage in major publications like ZDNet, VentureBeat,
and by Forbes staff writers since the last . It has become notable at this point
so is a even if it is a passing fad, as per WP:DEGRADE . Chagropango ( talk )
18:50, 5 July 2023 (UTC) [ reply ] Comment "A might have been a viable option
even 2 years ago, but considering that this platform is owned by HTC Corporation,
which is striving to build something on the level of the Facebook''s Meta, and
has garnered media attention in recent months (Probably also in Chinese from the
Taiwan media), I''m more inclined to retain and enhance the page." I’d rather
ask the editors from Taiwan or those competent in Chinese to double check the
media coverage in the manufacturer’s original language as well. -- Onetimememorial
( talk ) 20:32, 5 July 2023 (UTC) [ reply ] Relisted to generate a more thorough
discussion and clearer consensus. Please add new comments below this notice. Thanks,
L iz Read! Talk! 05:25, 12 July 2023 (UTC) [ reply ] Relisted to generate a more
thorough discussion and clearer consensus. Please add new comments below this
notice. Thanks, ✗ plicit 12:20, 19 July 2023 (UTC) [ reply ] Strong . With >20
cites from across five different years , why is this even up for consideration?
It''s a perfectly valid article with a good structure and lots of room for improvement
in future. Last1in ( talk ) 12:44, 19 July 2023 (UTC) [ reply ] with HTC Vive
. It seems to have made a splash in headlines, but contrary to what Last1in claims,
there is not coverage across "five different years", with only one article being
written in 2023 and the remainder being written in 2022. I''m unconvinced of any
lasting notability. SWinxy ( talk ) 00:47, 20 July 2023 (UTC) [ reply ] I am not
convinced of lasting notability, either, but my precognition is acting up . As
for sources, were you perhaps looking at retrieval date? Publish date shows five
years. 2016: VentureBeat; 2019: VentureBeat; 2021: Frater, Patrick (Variety);
2022: Auganix, James Dargan (MVI), Ochanji Sam (VRTimes), others; and 2023: Grant,
Rob (VRTimes), VentureBeat. I was unable to find an actual publication date on
a dozen more of them. Cheers, Last1in ( talk ) 14:47, 20 July 2023 (UTC) [ reply
] The 2016 and 2019 VentureBeat articles make no mention of Viverse. One is on
an app store and the other is on an app subscription package. SWinxy ( talk )
20:52, 20 July 2023 (UTC) [ reply ] . The platform is a big enough phenomenon
that even indirectly related events have gotten the attention of the VR media:
[15] , [16] , [17] . Deckkohl ( talk ) 17:24, 21 July 2023 (UTC) [ reply ] The
above discussion is preserved as an archive of the debate. Please do not modify
it. Subsequent comments should be made on the appropriate discussion page (such
as the article''s talk page or in a deletion review ). No further edits should
be made to this page.'
inference: true
---
# SetFit with BAAI/bge-base-en-v1.5
This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification.
The model has been trained using an efficient few-shot learning technique that involves:
1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.
## Model Details
### Model Description
- **Model Type:** SetFit
- **Sentence Transformer body:** [BAAI/bge-base-en-v1.5](https://huggingface.co/BAAI/bge-base-en-v1.5)
- **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
- **Maximum Sequence Length:** 512 tokens
- **Number of Classes:** 8 classes
<!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)
- **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)
### Model Labels
| Label | Examples |
|:------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| 0 | <ul><li>'Artisteer: The only two secondary sources I could find were this and this , neither of which are reliable sources. HyperAccelerated ( talk ) 01:04, 5 March 2024 (UTC) The article was written by a user named "Artisteer", and their only contributions to Wikipedia were on this article. There may be a WP: COI , but given that their last edits were many years ago, I\'m not sure what can be done about that now. HyperAccelerated ( talk ) 01:05, 5 March 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Internet and Software . WC Quidditch ☎ ✎ 01:37, 5 March 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, L iz Read! Talk! 00:22, 12 March 2024 (UTC) I can only find some software blogs saying it was abandoned about 10 yrs ago, then this [19] , neither of which is RS. I don\'t see any reliable sources we\'d use. Oaktree b ( talk ) 02:00, 12 March 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>'The Volume of Self : No chart placement, no record of notability. Fails WP:NALBUM: "All articles on albums or other recordings should meet the basic criteria at the notability guidelines, with significant coverage in reliable sources that are independent of the subject." as ATD unlikely as band article is itself at AfD. Alexandermcnabb ( talk ) 15:53, 12 September 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Music and United Kingdom . Alexandermcnabb ( talk ) 15:53, 12 September 2023 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, L iz Read! Talk! 23:08, 19 September 2023 (UTC) [ reply ] The article about the album\'s author was soft d , but had a vote saying that there was "no coverage for this musical group". I assume the same logic can be applied to this article. The first reference provides a list of songs in this album, and the second appears to be an article about this album on Blabbermouth , which is a heavy metal news site. As far as I can tell, that is the only source I could find for either of these articles, so this article doesn\'t meet the WP:GNG or WP:NALBUM . The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>"Anubha Sourya Sarangi : Sources are mostly about individual movies without significant coverage of the actress herself. No real evidence of notability * Pppery * it has begun... 16:00, 10 April 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Actors and filmmakers , Women , and Odisha . Spiderone (Talk to Spider) 18:58, 10 April 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, L iz Read! Talk! 23:27, 17 April 2024 (UTC) [ reply ] per nominator. I cannot find sourcing to satisfy notability requirements. Open to re-evaluating if some are found. — Sirdog ( talk ) 05:47, 19 April 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li></ul> |
| 4 | <ul><li>"Korina Adamou: The subject has earned at least eight caps for the Cyprus women's national football team . I am unable to find sufficient in-depth coverage from third-party sources, failing WP:GNG . The most I found was this and this . JTtheOG ( talk ) 00:02, 25 August 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Sportspeople , Women , Football , and Cyprus . JTtheOG ( talk ) 00:02, 25 August 2023 (UTC) This discussion has been included in WikiProject Football 's list of association football-related deletions. Spiderone (Talk to Spider) 19:07, 25 August 2023 (UTC) [ reply ] as above. Giant Snowman 08:20, 27 August 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>"Munkir (TV series): — Saqib ( talk I contribs ) 16:16, 18 May 2024 (UTC) This discussion has been included in the list of Pakistan-related deletion discussions . — Saqib ( talk I contribs ) 16:16, 18 May 2024 (UTC) This discussion has been included in the list of Television-related deletion discussions . WC Quidditch ☎ ✎ 17:13, 18 May 2024 (UTC) [ reply ] to List_of_programs_broadcast_by_TV_One_(Pakistan)#Drama_series - My, oh my! (Mushy Yank) 14:19, 22 May 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Relisting comment: Nominator appears to have copied and pasted the nominating rationale for another rush of AfD nominations, despite the numerous times others have cautioned the nominator about making a lot of nominations in a rush, so I am copying and pasting this relist remark. Please add new comments below this notice. Thanks, Doczilla Ohhhhhh, no! 09:45, 26 May 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>"Milo Nqoro : JTtheOG ( talk ) 03:26, 28 May 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Sportspeople , Rugby union , and South Africa . JTtheOG ( talk ) 03:26, 28 May 2024 (UTC) GNG . is a suitable WP:ATD . Rugbyfan22 ( talk ) 18:22, 28 May 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li></ul> |
| 6 | <ul><li>"Stockton Rush : — Crumpled Fire • contribs • 23:09, 22 June 2023 (UTC) [ reply ] On the contrary, he does very much meet point 2. And with regards to point 3, I will contest that it weren’t his actions during the expedition that caused the mishap, so he’s not independently notable for it. Basically he was just one of the occupants, his role in the loss of Titan wasn’t substantial. Carbon case of what this policy was written for. T v x 1 23:23, 22 June 2023 (UTC) [5] , [6] , [7] Dr. Swag Lord ( talk ) 23:13, 22 June 2023 (UTC) . [ reply ] Not true. If he had been a notable businessman prior to his death, he would have already have had an article. There’s nothing spectacular about his business activities. T v x 1 23:19, 22 June 2023 (UTC) [ reply ] He is now the central character in a major international news story. People will remember this incident and the person responsible for years to come. I'd say his backstory, which had many elements that led up to the implosion of the Titan, is relevant for historical purposes. 96.241.148.20 ( talk ) 23:26, 22 June 2023 (UTC) — 96.241.148.20 ( talk ) has made few or no other edits outside this topic. [ reply ] The central character in ONE EVENT ! People like him is exactly what Wikipedia:BLP1E was written for. Nothing in your comment justifies his article. T v x 1 23:30, 22 June 2023 (UTC) [ reply ] That argument doesn't work. BLP1E isn't satisfied because the third criterion states it has to be insignificant, whereas this was significant. 2A00:23C6:B894:FA01:815C:3D36:1E41:964D ( talk ) 23:39, 22 June 2023 (UTC) — 2A00:23C6:B894:FA01:815C:3D36:1E41:964D ( talk ) has made few or no other edits outside this topic. [ reply ] No his role in the incident has to have been substantial, which is not the case. There is nothing that suggests that his piloting of the vessel caused the breakup. His role in the accident isn’t in any way more important than that of the other four occupants. T v x 1 23:44, 22 June 2023 (UTC) [ reply ] His piloting of the vessel doesn't need to be the cause of the sub breaking up in order for his role to be substantial - that's a subjective interpretation of what is meant by 'substantial'. And clearly his role is more important than the other four occupants - this incident revolves entirely around a sub that he and those working for him designed and built. He himself dismissed concerns about the safety of the design, and then he himself was piloting it. The incident doesn't even happen without his involvement, so his involement is objectively substantial. 176.254.143.249 ( talk ) 23:51, 22 June 2023 (UTC) — 176.254.143.249 ( talk ) has made few or no other edits outside this topic. [ reply ] It's a big personal conjecture on you part to state it would not have happened without him. The expedition was executed by a company, not a person. They could have done that with another CEO as well. Also there is confirmation that design errors by him caused this, your are making wild assumptions here. There's nothing here about that incident that wasn't already included elsewhere. This is largely a content fork . T v x 1 23:56, 22 June 2023 (UTC) [ reply ] First of all, he wasn’t the pilot. Second of all, he’s the CEO of the company and had final say in the design decisions of the sub, which ultimately led to a lack of proper engineering and safety standards that led to the implosion. 42.3.105.87 ( talk ) 23:53, 22 June 2023 (UTC) — 42.3.105.87 ( talk ) has made few or no other edits outside this topic. [ reply ] We don't know that yet. There has not been any confirmation that the break-up was the result of design flaws. And even if so, there were multiple engineers working for the company who all could be responsible. T v x 1 23:57, 22 June 2023 (UTC) 'John Hinckley Jr., for example, has a separate article because the single event he was associated with, the Reagan assassination attempt, was significant, and his role was both substantial and well documented.' This article clearly meets those conditions and there is a precedent. 176.254.143.249 ( talk ) 23:42, 22 June 2023 (UTC) — 176.254.143.249 ( talk ) has made few or no other edits outside this topic. [ reply ] No it doesn't. These case are not comparable. That assassination attempt was completely orchestrated and executed by Hinckley. It wouldn't have happened without him. In this case, there's nothing to suggest that the accident was caused by Rush's piloting. His role in the cause wasn't substantial at all. He was just as much an occupant and victim as the other four. T v x 1 23:52, 22 June 2023 (UTC) [ reply ] His piloting of the vessel doesn't need to be the cause of the sub breaking up in order for his role to be substantial. The Reagan assassination does not happen without Hinckley and the Titan incident does now happen without Rush. 176.254.143.249 ( talk ) 23:55, 22 June 2023 (UTC) — 176.254.143.249 ( talk ) has made few or no other edits outside this topic. [ reply ] That analogy is just not true. He wasn't needed at all for that submersible to be operated. T v x 1 23:59, 22 June 2023 (UTC) [ reply ] Strong . Definitely notable for more than one event, there are a lot of sources ranging from years ago to now easily found online. Icehax ( talk ) 23:16, 22 June 2023 (UTC) [ reply ] Then why was no one interested in writing an article for him prior to his death? T v x 1 23:18, 22 June 2023 (UTC) [ reply ] That was then; this is now. Though deceased, he is an internationally recognized figure following this news story that made headlines around the world. 96.241.148.20 ( talk ) 23:29, 22 June 2023 (UTC) — 96.241.148.20 ( talk ) has made few or no other edits outside this topic. [ reply ] ONE news story. See WP:NOTNEWS and WP:BLP1E . Please familiarize yourself with Wikipedia’s policies before taking part in a procedure like AFD. T v x 1 23:34, 22 June 2023 (UTC) [ reply ] Strong . As per above. Death Editor 2 ( talk ) 23:21, 22 June 2023 (UTC) [ reply ] None of which are valid arguments, so neither is yours. T v x 1 23:24, 22 June 2023 (UTC) [ reply ] Nah, you are wrong in this case. Death Editor 2 ( talk ) 23:30, 22 June 2023 (UTC) [ reply ] No, you are. Please read the sites content policies. Anything worth mentioning about this person can be put in the articles on his company and on the accident that claimed his life. T v x 1 23:34, 22 June 2023 (UTC) [ reply ] . Clearly notable, and a valid split from OceanGate . I have to disagree with Tvx1 's assertion that If he had been a notable businessman prior to his death, he would have already have had an article. Given that we continue to write new articles on Wikipedia, that can't be the case. Mackensen (talk) 23:27, 22 June 2023 (UTC) 38, 22 June 2023 (UTC) 23C6:B894:FA01:815C:3D36:1E41:964D ( talk ) 23:42, 22 June 2023 (UTC) — 2A00:23C6:B894:FA01:815C:3D36:1E41:964D ( talk ) has made few or no other edits outside this topic. 06, 23 June 2023 (UTC) 29, 22 June 2023 (UTC) 31, 22 June 2023 (UTC) 43, 22 June 2023 (UTC) — 96.241.148.20 ( talk ) has made few or no other edits outside this topic. 07, 23 June 2023 (UTC) 38, 22 June 2023 (UTC) SIGCOV and WP:GNG . – Davey 2010 Talk 23:41, 22 June 2023 (UTC) 11, 23 June 2023 (UTC) 43, 22 June 2023 (UTC) 10, 23 June 2023 (UTC) 25, 23 June 2023 (UTC) 49, 22 June 2023 (UTC) 08, 23 June 2023 (UTC) 06, 23 June 2023 (UTC) 08, 23 June 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>'John Ross (blogger) : Sources in the article are mostly brief mentions in the context of his time as one of a largish group of advisors to Ken Livingstone, former mayor of London, and are therefore not WP:SIGCOV . The exceptions to this are op-ed pieces, interviews and commercial book-store websites (and therefore not reliable/independent). The Guardian "profile" is a single-sentence mention summing to 15 words. No instances of WP:SIGCOV found in my WP:BEFORE . WP:BLP article so should be based on high-quality sources. FOARP ( talk ) 13:21, 11 April 2023 (UTC) This discussion has been included in the list of Academics and educators-related deletion discussions . FOARP ( talk ) 13:21, 11 April 2023 (UTC) This discussion has been included in the list of United Kingdom-related deletion discussions . Shellwood ( talk ) 13:22, 11 April 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Authors , Politicians , Journalism , and China . TJMSmith ( talk ) 13:40, 11 April 2023 (UTC) [ reply ] There\'s a sportsperson with the same name, nothing found for this person. no sources found. Oaktree b ( talk ) 15:25, 11 April 2023 (UTC) AUTHOR case here. We need to look for further reviews for: China’s Great Road: Lessons for Marxist Theory and Socialist Practices (2021) [21] Thatcher and Friends: The Anatomy of the Tory Party (1983) The Great Chess Game (2016)[ [22] ] Don’t Misunderstand China’s Economy (date?) It\'s difficult because a number of relevant sources are likely to not be available online because they\'re from the early 80\'s or in Chinese. Jahaza ( talk ) 16:55, 11 April 2023 (UTC) [ reply ] It\'s also difficult because there\'s another John Ross, author of You Don\'t Know China , who has a web presence as a China expert. Jahaza ( talk ) 16:57, 11 April 2023 (UTC) [ reply ] Comment This [23] Evening Standard article, was actually published in October 2007 and is SIGCOV. Jahaza ( talk ) 17:02, 11 April 2023 (UTC) The Ups and Downs of Ken Livingstone (you can see some of that content here [24] ). That plus the Evening Standard article, plus the book reviews already identified, plus this in the Sunday Telegraph [25] . (There are also other less useful sources that document smaller facts and opinions [26] [27] [28] [29] [30] [31] [32] ) Jahaza ( talk ) 17:22, 11 April 2023 (UTC) NAUTHOR pass but I\'m OK to withdraw based on the Evening Standard and book coverage. There\'s already a ! vote on the board so I can\'t withdraw at this point though unless Oaktree b withdraws - what do you think Oaktree? FOARP ( talk ) 08:05, 12 April 2023 (UTC) [ reply ] I\'m rescinding my vote above. I\'m ok if it gets kept, with the new sources, as above. Oaktree b ( talk ) 14:51, 12 April 2023 (UTC) [ reply ] Thanks Oaktree b . In our defence, the refs in the article are bad and the many other John Rosses out there complicated performing a WP:BEFORE . FOARP ( talk ) 07:54, 13 April 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>"Jacob Dahl Jurgensen: The four External links are to websites that describe the artist's works, but no indepth content about the artist. After searching, unable to find sources to provide sufficient coverage. Created on 12 September 2007 JoeNMLC ( talk ) 05:16, 14 September 2023 (UTC) [ reply ] Withdrawn by nominator - article now has sufficient references to establish notability. Thankyou for improving this one. JoeNMLC ( talk ) 12:54, 17 September 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Artists and Denmark . Hey man im josh ( talk ) 11:59, 14 September 2023 (UTC) [ reply ] Comment I updated the article. WP:BEFORE shows he is in the British Museum collection. That makes him notable. Slim article, but I think it is a . -- WomenArtistUpdates ( talk ) 23:32, 16 September 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li></ul> |
| 1 | <ul><li>'Guy Protheroe : I\'m surprised to find so little, especially for an article this extensive, but I do not see evidence of notability here. The removal suggests sources in Google Scholar which may be of use here, but I could only find passing mentions in there as well so I\'m doubtful. QuietHere ( talk | contributions ) 03:41, 23 October 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Bands and musicians and England . QuietHere ( talk | contributions ) 03:41, 23 October 2023 (UTC) [ reply ] There are bios available, [14] and [15] , and some coverage in the NY Times [16] . He appears to have a chapter in this 1975 book but I can\'t access the text: [17] . Judging by his AllMusic credits, link , Protheroe appears to have an extensive career. Given this I have a suspicion that Protheroe may prove notable - but further work, potentially with offline sources, is needed to find the evidence. If kept, the article really does need an ounce or two of WP:TNT . Resonant Dis tor tion 20:29, 23 October 2023 (UTC) [ reply ] Internet Archive has British Music Now . Protheroe contributed a chapter (on Alexander Goehr ) to the book; except for a short "contributor" bio (likely autobiographical) he\'s not profiled in it. Jfire ( talk ) 03:42, 24 October 2023 (UTC) Clearly a notable (GNG) individual who is featured in multiple publications. Google Books , Google Ngram and Google Scholar . For these reasons, the article should be kept. Aye, the article clearly requires a clean up and referencing, but that\'s not a reason to the article. IJA ( talk ) 22:23, 25 October 2023 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, Daniel ( talk ) 11:37, 30 October 2023 (UTC) though simply pointing to google books and google scholar results is insufficient to demonstrate notability, I think the Milken Archive bio and couple of paragraphs in the NYT pointed to by ResonantDistortion are just about sufficient on their own to demonstrate a GNG pass. (The English Chamber Choir bio isn\'t helpful, as it is not independent, and as Jfire says Protheroe is a contributor to, not a subject of, British Music Now ). Caeciliusinhorto-public ( talk ) 14:52, 2 November 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>'Michael O\'Brien (New Hampshire politician): X ( talk ) 08:29, 18 February 2024 (UTC) This discussion has been included in the list of People-related deletion discussions . CAPTAIN RAJU (T) 08:44, 18 February 2024 (UTC) This discussion has been included in the list of New Hampshire-related deletion discussions . CAPTAIN RAJU (T) 08:44, 18 February 2024 (UTC) This discussion has been included in the list of Politicians-related deletion discussions . CAPTAIN RAJU (T) 08:44, 18 February 2024 (UTC) [ reply ] gets a freebie per NPOL. Djflem ( talk ) 11:58, 18 February 2024 (UTC) NPOL and there is information that he is from reliable sources , [44] [45] [46] (although I have to say they were far too hard to find and if he wasn\'t in the NH house it would be an easy ) Shaws username . talk . 14:24, 18 February 2024 (UTC) [ reply ] -- Clearcut case per NPOL. Central and Adams ( talk ) 16:41, 18 February 2024 (UTC) [ reply ] . Meets WP:NPOL . He is a member of the New Hampshire House of Representatives . MoviesandTelevisionFan ( talk ) 19:11, 18 February 2024 (UTC) [ reply ] per previous votes. BottleOfChocolateMilk ( talk ) 23:36, 18 February 2024 (UTC) [ reply ] Hold our horses - NPOL exists because there\'s a presumption of coverage. New Hampshire has one legislator for every 3,300 people, which is one of the lowest if not the lowest in the entire world. Are we sure he has coverage as a result? There\'s nothing in the article apart from his legislative profile and the sources found here aren\'t about him. SportingFlyer T · C 10:37, 19 February 2024 (UTC) [ reply ] It can\'t be true that NPOL exists because there\'s a presumption of coverage. If there were coverage the GNG would suffice. NPOL must exist because sometimes there\'s not coverage but the subject is notable anyway. The presumption of notability in NPOL is not a rebuttable presumption . It\'s a guarantee of notability. That this is the case is clear from the discussion of local officials and unelected candidates at the bottom of the guideline, where it states: Just being an elected local official, or an unelected candidate for political office, does not guarantee notability If elected state level officials weren\'t guaranteed notability by NPOL it wouldn\'t be necessary to explicitly state that local officials were not. Central and Adams ( talk ) 13:43, 19 February 2024 (UTC) [ reply ] That\'s not what "presumed" means - the dictionary definition is literally "to take for granted as being true in the absence of proof to the contrary." We typically presume politicians will be notable because they should easily have received significant coverage, even if we can\'t find coverage of them. I\'m not convinced that\'s the case here. SportingFlyer T · C 21:41, 19 February 2024 (UTC) [ reply ] It literally is what one sense of the word "presume" is. I don\'t know what dictionary you\'re using, but the OED notes what you quoted as only one of two meanings, the other being "To assume; to take for granted; to presuppose; to anticipate, count upon, or expect". Which sense is intended is, as with all polysemous words, determined by context, which is why I argued from the context that the meaning here is as I stated. You\'re talking about the definition of a rebuttable presumption. My argument is that the sense meant here is an irrebuttable presumption. Central and Adams ( talk ) 22:24, 19 February 2024 (UTC) [ reply ] @ Central and Adams Greetings. You said, NPOL must exist because sometimes there\'s not coverage but the subject is notable anyway . -If a person is indeed notable, then why won\'t they get proper coverage? To my understanding, notability is vehemently based on coverage. How can a person be notable and not have coverage? What is notability based on if not on coverage? We give people freebies if they\'ve won some reputable awards or an academic or a notable politician, but I don\'t think you\'d find many person not having coverage meeting any of these criteria properly. I\'d like your take on that. (PS: Agreeing with @ SportingFlyer on this) We typically presume politicians will be notable because they should easily have received significant coverage, even if we can\'t find coverage of them. I\'m not convinced that\'s the case here. . X ( talk ) 08:20, 21 February 2024 (UTC) [ reply ] If notability is based solely on coverage then why would we need any notability guidelines other than the GNG? The very fact that NPOL contains a presumption of notability shows that it must apply in the absence of coverage. If there were coverage it wouldn\'t be necessary to have a presumption of notability. If all NPOL meant were that politicians are notable unless there\'s no coverage it would say the same thing as the GNG, so why would it exist? Central and Adams ( talk ) 09:48, 21 February 2024 (UTC) [ reply ] We really only do have the GNG. We have presumptions because it makes it easier to figure out what can and can\'t be covered, but in the past few years we have generally tied the presumptions very close to the GNG. The NPOL presumption exists because if you\'re a member of a state legislature, it is almost certain you will have been written about in reliable secondary sources, which is helpful for say someone who was a member of a historical legislature who we can\'t access sources for. In O\'Brien\'s case, he\'s an active legislator, but one source is just the state website, the other source in the article just shows he\'s an alderman (the only thing on him on that website is his address) and a ballotpedia page, which is a wiki. Because of the fact there\'s no secondary information we can use to build this page out, and also due to WP:BLP concerns, it\'s probably best if this were ed somewhere, and the information in the article d there until we can write a stand-alone article. Also, I think this would be pretty close to being New Hampshire-specific, considering how few people vote on state legislators there. Most other MPs have much larger constituencies and as such have much more written about them. SportingFlyer T · C 16:21, 21 February 2024 (UTC) [ reply ] Not sure why you say there\'s no secondary information. It took me about ten minutes to find a ton of it, which I added to the article. But I still maintain that even if this weren\'t possible to do the dude would still pass NPOL. Also, it\'s not true that we really only do have the GNG. WP:N says explicitly that a subject is notable if it passes either the GNG or a subject specific guideline. Central and Adams ( talk ) 17:03, 21 February 2024 (UTC) [ reply ] That\'s not how it\'s been interpreted recently for most guidelines anymore, but there are a few exceptions. The new coverage removes any objections I\'ve had, though. SportingFlyer T · C 17:08, 21 February 2024 (UTC) NPOL . As long as it is verifiable the individual holds the office, state legislators are worthy of a stand alone article. There will be information about the subject in the official pages of the NH legislature - including votes taken and bills introduced and sponsored. There are records of election results. All of this is good, verifiable information that can build a strong article, even if there is limited newspaper coverage now. -- Enos733 ( talk ) 19:50, 19 February 2024 (UTC) Clearly meets WP:NPOL , as others have mentioned. Hey man im josh ( talk ) 19:12, 20 February 2024 (UTC) NPOL . If some believe GNG isn\'t met, then NPOL\'s validity must be challenged to the community as a whole like WP:NSPORTS was. Best, GPL93 ( talk ) 19:12, 21 February 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>'Man-Ching Donald Yu: 日期20220626 ( talk ) 01:18, 9 September 2023 (UTC) [ reply ] It\'s unclear whether Authority control can demonstrate his notability, but it seems that there are fewer links within Authority control. 日期20220626 ( talk ) 01:20, 9 September 2023 (UTC) [ reply ] Comment . The article includes several references. As well, the references in the corresponding Chinese article at zh:余文正 may be helpful. See also the previous AFD at Wikipedia:Articles for deletion/Man-Ching Donald Yu Eastmain ( talk • contribs ) 01:52, 9 September 2023 (UTC) 余文正 provide information about him, but they are actually from personal websites. When I searched for his name on Google, I couldn\'t find any satisfactory results. 日期20220626 ( talk ) 03:14, 9 September 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Bands and musicians and Hong Kong . Eastmain ( talk • contribs ) 01:59, 9 September 2023 (UTC) [ reply ] Comment , there are some pointers, but hard to tell right now. Contemporary composer, very niche. Here\'s an album of compositions [24] Ralph P Locke, in Music and Letters , Volume 102, Issue 3, August 2021, Pages 641–643, describes him as "a noted composer as well as scholar". Via TWL [25] — siro χ o 03:05, 9 September 2023 (UTC) [ reply ] Music and Letters only mentioned his name briefly. 日期20220626 ( talk ) 03:18, 9 September 2023 (UTC) [ reply ] per the significant coverage in multiple independent reliable sources . The subject passes Wikipedia:Notability (people)#Basic criteria , which says: People are presumed notable if they have received significant coverage in multiple published secondary sources that are reliable , intellectually independent of each other, and independent of the subject . If the depth of coverage in any given source is not substantial, then multiple independent sources may be combined to demonstrate notability; trivial coverage of a subject by secondary sources is not usually sufficient to establish notability. Sources Canfield, David DeBoor; Nockin, Maria (March–April 2013). "Yu Symphony No. 1. From the Depth. Octet for Strings. Sunset in my Homeland. The Maximum Speed of Raphael\'s Madonna. Explosion for Piano. Disintegration for Piano and Electronics. Two Poems by Ya Hsien. Breeze". Fanfare . Vol. 36, no. 4. pp. 172–174. ProQuest 1287039828 . David DeBoor Canfield wrote: "The present CD contains a generous sampling of music by Hong Kong composer and pianist, Dr. Man-Ching Donald Yu, who was born in 1980. As a pianist, Yu made his debut at the age of 16 with the Pan Asia Symphony Orchestra, and eventually earned a B.A. degree from Baylor University. Further musical studies took him to the Internationale Sommerakademie Universität Mozarteum in Salzburg, and he completed his education, being awarded a Ph. D. in composition and music theory at Hong Kong Baptist University. He is currently on the faculty of the Hong Kong Institute of Education. The more than 150 compositions in Yu’s portfolio range from instrumental, vocal, and chamber pieces to large-scale operatic, choral, and symphonic works. The music on this, the second CD devoted to the composer’s music, has been selected to give an overview of the breadth of the genres in which this composer writes." Maria Nockin wrote: "Man-Ching Donald Yu is an intriguing composer who writes in several different styles. The first work on this disc is his First Symphony which has three movements that are grouped together on one 20-minute track. The first movement is something of a prelude to the stronger and darker music to come. There is a great deal of melodic material, especially for the Lugansk orchestra’s brass section. It is buoyed up by the strings and punctuated by gestures from the percussionists." Hinterbichler, Karl (July 2010). "Man-Ching (Donald) Yu. Solemn Elegy for four trombones. Orlando, FL: Wehr\'s Music House, 2007. Playing time 2:00. Score and parts". ITA Journal . 38 (3). International Trombone Association : 46. ProQuest 748815971 . The article notes: "Donald Yu was born in 1980 in Hong Kong. He earned a Bachelor of Music degree from Baylor University, where he studied piano, composition and conducting and pursued further studies in Austria, Italy, and Germany. He has composed over 100 works in a variety of media. A number of these have been published and recorded. In addition to Solemn Elegy he has composed two other works for trombone, Reflections for Trombone Choir (2006) The Refraction of Shadow for trombone and piano (2007). He is currently a Ph. D. candidate in Composition at Hong Kong Baptist University. Solemn Elegy is traditional harmonically and melodically. It contains no musical or technical difficulties that could not be surmounted by an average college quartet. The title describes its musical qualities quite well. If you need a slow, short, somber work to fill out a program, this fits the bill." Rees, Carla (Fall 2014). "New Music Without Borders, Volume 2". The Flutist Quarterly . Vol. 40, no. 1. National Flute Association . p. 71. ProQuest 1619361371 . The review notes: "Man-Ching Donald Yu\'s contribution, "Breeze," is intended for "young performers and children who have studied the instrument for a short time and have not been exposed to contemporary music. " Based on five chromatic pitches, this three-minute piece uses key clicks, flutter-tonguing, pizzicato, whistle tones, jet whistles, and glissandi to create an evocative sound. Another short piece, Fernando Maglia\'s "Tropos II," uses similar techniques in a slightly more complex rhythmic and harmonic language, providing a useful progression for students. Attention to detail can be developed with this piece, given its frequent dynamic changes and contrasting moods between short phrases." Camilleri, Silvio John (2013-04-14). "An ambitious effort that delighted" . Times of Malta . Archived from the original on 2023-09-10 . Retrieved 2023-09-10 . The article notes: "The second item was the world premiere for Sign of Spring composed in 2012 by Man-Ching Yu. Born in Hong Kong, this composer is frequently inspired by paintings. The work is tinged with an Oriental touch, which intermingles with impressionistic, Western elements; a bit like a Toru Takemitsu composition." Walker, Brian (October 2013). "Music Review: "Fishing in Snow," by Man-Ching (Donald) Yu". International Trumpet Guild Journal . Vol. 38, no. 1. International Trumpet Guild . p. 98. ProQuest 1487696682 . The abstract notes: "Walker reviews a composition by Yu for trumpet and piano (MusicaNeo)." https://www.manchingdonaldyu.com/reviews Internet Archive has additional reviews. There is sufficient coverage in reliable sources to allow Man-Ching Donald Yu ( Chinese : 余文正 ) to pass Wikipedia:Notability#General notability guideline , which requires "significant coverage in reliable sources that are independent of the subject". Cunard ( talk ) 08:31, 10 September 2023 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Relisting comment: A source analysis of new sources would be welcome. Please add new comments below this notice. Thanks, L iz Read! Talk! 01:16, 16 September 2023 (UTC) BASIC , met per sources found by Cunard. — siro χ o 02:56, 16 September 2023 (UTC) [ reply ] per the sources found by Cunard. Mccapra ( talk ) 06:06, 16 September 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li></ul> |
| 2 | <ul><li>"Manning Community School District: PaulGamerBoy360 ( talk ) 03:22, 21 July 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Education , Schools , United States of America , and Iowa . PaulGamerBoy360 ( talk ) 03:22, 21 July 2023 (UTC) [ reply ] The navbox found in the article is full of redlinks. That suggests a topic area needing expansion, not scaling back. The nominator's rationale is misleading in that the article is orange-tagged for needing expansion, not for lacking sources. Anyway, to IKM–Manning Community School District . The article on the other pre-r component of that district is also light on third-party sources. RadioKAOS / Talk to me, Billy / Transmissions 04:51, 21 July 2023 (UTC) I will check Newspapers.com to see how many sources I can find about the pre-r district. WhisperToMe ( talk ) 11:41, 21 July 2023 (UTC) I found several newspaper article sources about how the creation of the district was legally disputed in court by another school district (the Iowa Supreme Court ultimately upheld the creation of this district). Additionally there was a legal dispute in regards to two areas being moved into this districgt. This is certainly not routine coverage by any stretch of the definition, and so this should secure notability of this topic. WhisperToMe ( talk ) 12:01, 21 July 2023 (UTC) [ reply ] - to IKM–Manning Community School District . Historic predecessor districts should be covered in the current district's article unless WP:FORK becomes a concern. Even with additional content, size is not a concern here. The other predecessor district's articles should likewise be d, possibly dividing the history section at the target article into subsections to do so. By all means, this title should remain as a . 4.37.252.50 ( talk ) 01:50, 22 July 2023 (UTC) WEIGHT becomes too much for a particular section (for example, the weight of the information about the former Manning district, which operated from 1959 to 2011), then that former district should have its own article. I'm still finding content about the 1959-2011 period, and I think that there may be enough for this district to have its own article. Also there is notability by being a populated, legally recognized place (as per Wikipedia:Notability_(geographic_features)#Settlements_and_administrative_regions ). Various former municipalities in Japan, which have since d into larger ones, would count as being legally recognized places. Former school districts are also legally recognized places. WhisperToMe ( talk ) 22:53, 22 July 2023 (UTC) 14, 23 July 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>"Trikut cable car accident: LibStar ( talk ) 04:05, 31 August 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Events , Transportation , and Jharkhand . Spiderone (Talk to Spider) 08:32, 31 August 2023 (UTC) [ reply ] or to a new section about the cablecar on the Trikut Hill article. There is plenty of more recent coverage available, e.g. [41] , [42] , [43] , [44] , [45] . The most recent of those is from less than a week ago, so the nominator has clearly not done a (sufficient) WP:BEFORE . I'm happy with either ing or merging, but there is definitely no cause for deletion Thryduulf ( talk ) 09:42, 31 August 2023 (UTC) [ reply ] I have now expanded it based on two of those sources but more improvement is certainly possible. Thryduulf ( talk ) 10:07, 31 August 2023 (UTC) [ reply ] to Trikut Hill. Not worthy of its own article per WP:NOTNEWS . sixty nine • whaddya want? • 18:03, 31 August 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>"Sound BlasterAxx: ~ T P W 15:32, 28 June 2023 (UTC) This discussion has been included in the list of Technology-related deletion discussions . ~ T P W 15:32, 28 June 2023 (UTC) This discussion has been included in the list of Products-related deletion discussions . Spiderone (Talk to Spider) 19:31, 28 June 2023 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Relisting comment: Ineligible for soft deletion. Please add new comments below this notice. Thanks, ✗ plicit 23:39, 5 July 2023 (UTC) [ reply ] per nominator. Performing a WP:BEFORE search returns no RS . FatalFit | ✉ | ✓ 00:01, 6 July 2023 (UTC) [ reply ] Comment It does get some reviews in bigger technical websites [20] [21] [22] [23] [24] [25] but there's very little here beyond run-of-the-mill coverage and no particular reason to the current content (which is largely copied from datasheets and press releases) as a separate article. -- Colapeninsula ( talk ) 11:04, 6 July 2023 (UTC) [ reply ] per nomination, reviews found by Colapeninsula are good, but as they say, not evidence of separable notability. — Ganesha811 ( talk ) 14:18, 6 July 2023 (UTC) [ reply ] Weak , otherwise per nom. IMO the third-party reviews cited in this AFD and the previous one are sufficient to pass the GNG (edit: and also WP:NCORP / WP:PRODUCTREV , given their significance, depth and apparent independence). However, given particular problems of this article and the general problems of product articles, I would ordinarily still favor a . But the problem IMO is that Sound Blaster is already rather unwieldy and is getting into WP:SIZESPLIT territory. It seems like merging now is just going to make more work for future splitters, with no real benefit. (I wonder, though, if perhaps some sort of reconfiguration into a List of Sound Blaster USB products or some such, spinning off that entire L3 heading from Sound Blaster , might be better way to structure our coverage.) -- Visviva ( talk ) 22:30, 6 July 2023 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, L iz Read! Talk! 23:21, 12 July 2023 (UTC) [ reply ] , agreeing with the assessment of potential sources by Colapeninsula that they are run-of-the-mill even if reliable and in-depth. SWinxy ( talk ) 23:33, 19 July 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li></ul> |
| 3 | <ul><li>'A-Plus (rapper): Just because we have several articles about music produced by him does not make him notable, I find that he is not notable as a musician or a producer. Nagol0929 ( talk ) 15:59, 4 April 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Bands and musicians and Music . Nagol0929 ( talk ) 15:59, 4 April 2024 (UTC) I haven\'t looked closely yet as to whether his article deserves to stay, but it seems to me a to Souls of Mischief might be a better option than outright deletion... yes, I know he is part of Hieroglyphics (group) as well and therefore WP:XY may be considered here, but Hieroglyphics is all of Souls of Michief plus four other people, so he\'s still a part of Hieroglyphics as a member of Souls of Mischief. Richard3120 ( talk ) 16:13, 4 April 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: California and Colorado . WC Quidditch ☎ ✎ 18:51, 4 April 2024 (UTC) He clearly passes WP:NMUSIC#C6 if he\'s part of two notable production groups. That doesn\'t mean we have to have a standalone article on him, just noting a discrepancy in the nom statement. Mach61 20:25, 4 April 2024 (UTC) 42, 4 April 2024 (UTC) [ reply ] Weak Other than the 2 sources provided by above editor, there are not enough reliable coverage and 2 of the sources are interviews. Bradelykooper ( talk ) 08:34, 10 April 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, Shadow311 ( talk ) 16:03, 11 April 2024 (UTC) [ reply ] . None of the sources appear to be reliable, but a search of his name would go to the band\'s article, a compromise that we do sometimes. Bearian ( talk ) 14:36, 15 April 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, Shadow311 ( talk ) 18:42, 18 April 2024 (UTC) AllMusic is a reliable source as per [ [108] ] and the bio and album review are not interviews as someone else claimed, Atlantic306 ( talk ) 22:54, 18 April 2024 (UTC) the only problem is that AllMusic isn’t being used as a reference and all 3 of the references are interviews. Of those only 1 is about A-Plus. Nagol0929 ( talk ) 03:26, 19 April 2024 (UTC) 00, 19 April 2024 (UTC) [ reply ] There\'s a ton of sourcing (yes, from reliable sources) available on this guy in Google News and Books searches, over a period of decades. It\'s true that most of them are brief mentions, but with all of the info available, surely the article could be built out and sourced better than it is now. I had to get a little creative in looking for sources since "A plus" is such a generic term, but combining his name with "Hieroglyphics" or "Souls of Mischief" yields many good results. Fred Zepelin ( talk ) 19:32, 23 April 2024 (UTC) [ reply ] @ Fred Zepelin may you link said results? Mach61 01:40, 25 April 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Relisting comment: Final relist. Please add new comments below this notice. Thanks, Daniel ( talk ) 05:11, 26 April 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>"Owen Buckley : All I found were transactional announcements ( 1 , 2 ) with a combined five-ish sentences of independent coverage. JTtheOG ( talk ) 16:45, 30 April 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Sportspeople , Rugby league , and England . JTtheOG ( talk ) 16:45, 30 April 2024 (UTC) Sufficient room for expansion, but not enough coverage in current state. Mn1548 ( talk ) 13:04, 1 May 2024 (UTC) [ reply ] Weak I was able to find this source, which I think is detailed enough be considered non-trivial, but it's a local media article, so it's pretty borderline. J Mo 101 ( talk ) 11:23, 2 May 2024 (UTC) [ reply ] Transactional announcements such as signings and trades are not considered in-depth sourcing, especially when most of it is in quotes. JTtheOG ( talk ) 18:15, 2 May 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, The Herald (Benison) ( talk ) 18:10, 7 May 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, L iz Read! Talk! 23:20, 14 May 2024 (UTC) [ reply ] Weak I lean to as not quite meeting notability guidelines but will support the consensus of the group of editors. Go4thProsper ( talk ) 18:38, 18 May 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>'Master Hilarion : But it\'s simply false that no reliable independent sources exist. Some are already listed on the page. The Masters Revealed: Madam Blavatsky and the Myth of the Great White Lodge and Radiance from Halcyon: A Utopian Experiment in Religion and Science are both academic studies that appear to have significant coverage of Hilarion. There\'s an entry in Encyclopedia of Occultism and Parapsychology , published by Gale Group . And, if you weed out the non-RSs from Google Scholar, you find plenty of modern scholarship that covers him -- [8] , [9] , [10] , [11] , and so on. Jfire ( talk ) 05:31, 3 January 2024 (UTC) This discussion has been included in the list of Religion-related deletion discussions . WC Quidditch ☎ ✎ 05:55, 3 January 2024 (UTC) [ reply ] per Jfire. Skyerise ( talk ) 10:44, 3 January 2024 (UTC) [ reply ] The article has not had any good academic or scholary references added to it in over a decade (in fact none since it was created). There are also now copyright issues. Neither of the books Jfire mentions have significant coverage of Master Hilarion they have a few scattered lines, the papers on brill.com do not mention Master Hilarion in any detail. One of the sources Jfire lists as "modern scholarship" covering Master Hilarion is this paper on the gay activism of Wallace de Ortega Maxey , it has a mere line about Master Hilarion [12] . Wallace de Ortega Maxey is a non-notable figure himself. I don\'t see how any this is relevant or will establish notability. We want in depth scholarly sources that mention this topic. There is no point in citing a paper just because it has one line about the subject. Anyone can look on Google Scholar, just because you get a few hits does not mean these sources contain significant coverage. If you look on JSTOR, the same thing happens. There is only passing mention of Master Hilarion [13] . This does not establish notability. In conclusion, only the Gale Group source was a useful one but a single source is not enough to build an article on. I see a serious lack of independent neutral sources on this topic. I vote . Psychologist Guy ( talk ) 18:13, 7 January 2024 (UTC) [ reply ] This is an inaccurate summary of the sources. In The Masters Revealed , Hilarion is the primary topic of pages 59–62. Here\'s the first paragraph from these pages: ONE OF THE MORE ELUSIVE MASTERS of HPB\'s Egyptian Brotherhood is the man she called Hilarion (or Illarion) Smerdis. The authorship of several fictional works published by HPB has been attributed to him, including the stories "Unsolved Mysteries," "The Ensouled Violin," and "The Silent Brother." Along with Morya and Koot Hoomi, Hilarion has continued to be an alleged source for "channelers" in the twentieth century, most notably Canadian medium Maurice Cooke. In May 1875, HPB\'s scrapbook noted that Hilarion and a companion "passed thro\' New York & Boston, thence thro\' California and Japan back." In 1878, the same scrapbook, referring to a letter or psychic transmission received from Hilarion, noted "panic in England. Russians at Constantinople. Gorchakov hoodwinks Disraeli." This seems to indicate shared interests that are more political than spiritual. In July 1881, The Theosophist published Hilarion\'s report of his explorations of Zoroastrian ruins in Armenia. After the society moved to Adyar, Smerdis sent a letter advising Olcott that Serapis wanted him to travel in South India and Ceylon. Hilarion was described by HPB as a Greek gentleman with a black beard and long flowing white garments, looking from a distance like Serapis, and passing through Bombay en route to Tibet for his "final initiation." After going to Tibet, he allegedly inspired Mabel Collins\'s Idyll of the White Lotus and Light on the Path, although this was later denied by Collins. I don\'t have access to the full text of Radiance from Halcyon , but if anything its coverage of Hilarion appears to be even more significant than The Masters Revealed -- hits in 47 snippets, many of which are clearly discussing specific aspects of Theosophist beliefs about Hilarion. Here is what the paper on Maxey says of Hilarion: Maxey wrote extensively on the esoteric wisdom of Theosophy, tracing it through the avatar of the Master Hilarion, located by Maxey in various incarnations from Orpheus in 7000 BC through Ramses II, St. Paul, Montezuma, Hiawatha, and George Washington... Maxey also found Hilarion\'s work at play in the American Revolution, particularly in "the beautiful and occult vision which took place at Philadelphia," which he felt best embodied the "Universal Brotherhood unhampered by creed, race, or color." Here is what another of the papers says: Adepts and brothers were often experienced in their astral bodies. A May 1875 article in the Spiritual Scientist mentioned that one or more ‘Oriental Spiritualists of high rank’ had just arrived in the United States, whom Blavatsky identified as At[rya] and Ill[arion] passing through New York and Boston en route to California and Japan.43 Illarion (also called Hilarion), a Greek Cypriot adept, features as an elusive figure in Blavatsky’s memoirs. She had first met him on Cyprus in 1860 and again in Egypt in 1870. As a visitor to New York, he is supposed to be a physical body, but there are also indications that his astral body or projection is involved. She described Illarion with his ‘dark pale face, black beard and flowing white garments and fettah’ as ‘the form of a man’ whom Olcott and others met about their New York apartment, and she also referred to him as ‘John King’ because her companions might find it easier to accept a spirit than the astral body of a living man. Hilarion also collaborated with her in the writing of her occult stories and signed himself ‘Hilarion Smerdis’. This constitutes significant coverage, demonstrating in depth commentary and analysis of Hilarion\'s role in Theosophy. Jfire ( talk ) 19:28, 7 January 2024 (UTC) [ reply ] Comment This isn\'t in-depth commentary or significant coverage, only the first piece of green text you quote has some information so I agree that is a good source but it doesn\'t give us much else. These sources might be good if this article was a biography of Maxsey or Eugene O\'Neill but this article is about Master Hilarion. I don\'t see how a good article can be built by cherry-picking like this. Wallace de Ortega Maxey was a gay rights activist who is non-notable himself, I am not sure why he is relevant to an article on Hilarion. Why are we citing him? This source you cited from 1960 is on the Irish playwright Eugene O\'Neill [14] . It might be useful for his own biography or for a line of information, but it is not going to add significant coverage. This is not in-depth coverage from academics evaluating the Theosophical claims of Saint Hilarion, the last two are not strong sources and only have passing mention of Hilarion. This doesn\'t establish notability. If this topic was notable, historians would have written full papers on it. Psychologist Guy ( talk ) 19:54, 7 January 2024 (UTC) [ reply ] I\'m going to bow out after this and let others make their own assessment of these sources, but I just want to say that "historians have have written full papers on it" is not a Wikipedia notability criterion. We can and do cover many topics that don\'t have "full papers" written by historians. Jfire ( talk ) 20:16, 7 January 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, L iz Read! Talk! 08:41, 10 January 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Please add new comments below this notice. Thanks, Sandstein 14:03, 17 January 2024 (UTC) Whole article has been removed for a potential copyvio, likely not helping much at this point. Oaktree b ( talk ) 16:07, 17 January 2024 (UTC) BACKWARDSCOPY . Jfire ( talk ) 16:45, 17 January 2024 (UTC) [ reply ] Relisted to generate a more thorough discussion and clearer consensus. Relisting comment: Final relist. Please add new comments below this notice. Thanks, NotAGenious ( talk ) 18:52, 24 January 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li></ul> |
| 5 | <ul><li>'50.45.180.190 (Vandalism Bandit): I initially tagged as a WP:G3 but, upon reflection, it doesn\'t meet the definition. I can\'t see this meeting any of the speedy deletion criteria so I\'m sending it to AfD. The creator seems adamant that because the article is \'funny\', it doesn\'t need to meet our notability guidelines. I\'m hoping that common sense can prevail and WP:SNOW can happen. Spiderone (Talk to Spider) 23:10, 18 October 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Internet and Oregon . Spiderone (Talk to Spider) 23:10, 18 October 2023 (UTC) [ reply ] An admin really should just come along and it. There\'s no use in holding an AFD for an article that should be speedy d but can\'t because it hardly fits under any criteria (though maybe it fits under A11, since there\'s no such thing as a "vandalism bandit"). Waddles 🗩 🖉 23:35, 18 October 2023 (UTC) [ reply ] as clearly unfit for an encyclopedia. Completely non-notable. Schminnte [ talk to me ] 23:56, 18 October 2023 (UTC) A7 or something. Skynxnex ( talk ) 00:26, 19 October 2023 (UTC) [ reply ] Speedy . Misplaced humorous page. The author of this article, @ MrHistoryH also looks NOTHERE. I did laugh with this article, though. SparklyNights 03:21, 19 October 2023 (UTC) per WP:A11 . Not a notable topic. User:Let\'srun 01:35, 19 October 2023 (UTC) [ reply ] Strong . Per all above. Not a humor page. This straight up looks like an attack page against a school. 🛧 Midori No Sora♪ 🛪 ( ☁＝☁＝✈ ) 06:22, 19 October 2023 (UTC) A7 and/or WP:A11 per the above comments. If not, we may have to wait for the snow to fall... Spiderone (Talk to Spider) 08:33, 19 October 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>"Maarten Nagtegaal: Fram ( talk ) 13:30, 30 May 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: People and Netherlands . Fram ( talk ) 13:30, 30 May 2023 (UTC) [ reply ] Non-notable individual, there is zero coverage for him. He appears to be the son of a rich person, which isn't enough for notability. Oaktree b ( talk ) 13:34, 30 May 2023 (UTC) [ reply ] . Completely irrelevant non-notable figure. ULPS ( talk ) 14:42, 30 May 2023 (UTC) [ reply ] , non-notable figure. Creator claims that the information comes from knowing him personally (see talk). Schminnte ( talk • contribs ) 16:05, 30 May 2023 (UTC) [ reply ] Comment . This might be a different person with the same name. Also probably a different person: This . The corresponding article in the Dutch Wikipedia has been d. Twice. Eastmain ( talk • contribs ) 22:49, 30 May 2023 (UTC) [ reply ] almost a speedy for an unreferenced BLP. LibStar ( talk ) 01:35, 31 May 2023 (UTC) [ reply ] Speedy since cyber harassment of a lving person. Admin assistance appreciated! gidonb ( talk ) 19:16, 1 June 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>"XinFin : Sourcing consists of press release reprints or sources with only trivial mentions of XinFin. ~ A412 talk! 19:56, 15 February 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Cryptocurrency , Finance , Companies , Software , and Singapore . ~ A412 talk! 19:56, 15 February 2024 (UTC) [ reply ] Feel free to remove the article. S from the Rebel Moon ( talk ) 20:27, 15 February 2024 (UTC) [ reply ] Yes, I agree with your point of view. S from the Rebel Moon ( talk ) 20:29, 15 February 2024 (UTC) 48, 15 February 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li></ul> |
| 7 | <ul><li>"Motherless (disambiguation): A disambiguation page is not required ( WP:ONEOTHER ); the primary topic article has a hatnote to the only other use. Where the primary topic Motherless should target ( Single parent or Orphan ) is a matter for WP:RFD and does not require a disambiguation page. Shhhnotsoloud ( talk ) 09:26, 20 May 2023 (UTC) [ reply ] Withdrawn by nominator with thanks to other editors for finding other entries for the page. Shhhnotsoloud ( talk ) 06:54, 26 May 2023 (UTC) This discussion has been included in the list of Disambiguations-related deletion discussions . Shhhnotsoloud ( talk ) 09:26, 20 May 2023 (UTC) [ reply ] Agreed , certainly don't need this disambiguation page because at best it refers to only two topics. I don't think it needs to refer to Single parent anyway, as that article is looking at things from the perspective of the parent, not the state of the child, and if a link to Orphan is appropriate, that can be done by a hat-note. If more topics appear, then of course disambig can be reinstated. Elemimele ( talk ) 13:07, 20 May 2023 (UTC) [ reply ] Postpone this discussion until an RfD has decided where to point the Motherless Primary Topuc (I support Orphan as better than Single parent ). Then we can agree to this dab page, once we know where to put the hatnote which replaces it. Pam D 22:05, 21 May 2023 (UTC) [ reply ] Postpone I've added three potential articles, and I'm sure our descendents will find many more. No Swan So Fine ( talk ) 08:25, 23 May 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li><li>'Irmulco, California: This location is non-notable; coordinates given lccate to empty forest. This is one of dozens of mass-created stubs on nonexistent California locations created by the same editor during a short period in 2009, based only on GNIS coordinates. The fact that there was once a post office by this name does not establish notability; this was likely just a temporary logging camp. WeirdNAnnoyed ( talk ) 18:46, 24 October 2023 (UTC) I\'m going to withdraw my recommendation that the article be d. Another user has added some reasonably good sourcing and I think the article now has enough to . The article\'s current state is exactly how articles about small vanished communities should be...i.e., not just "Xyz is a location at zzz coordinates, and it had a post office in 1858". WeirdNAnnoyed ( talk ) 02:18, 25 October 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Geography and California . WeirdNAnnoyed ( talk ) 18:46, 24 October 2023 (UTC) [ reply ] comment This an example of place which was definitely "there" in some sense but about which we can\'t really say anything definite. We can\'t even really characterize it well. I\'ve found one reference to a public school there, and another passing reference to people living there, but as the nom here says, the name and location tends to suggest it was a logging camp of perhaps greater than usual permanence. There\'s some possibility that a history of the logging railroad might have more information, but without that it\'s hard to defend ing this. Mangoe ( talk ) 19:37, 24 October 2023 (UTC) [ reply ] like many places in the American west, it was a settlement for as long as the mineral seam lasted, or the lumber mill was hiring, etc, and then when the industrial or commercial interest checked out, it faded away, but it was a substantial settlement for a time in a thinly populated part of the world. It\'s part of the answer to the question "where did the redwoods go?" and I think it\'s notable enough to stay. jengod ( talk ) 23:44, 24 October 2023 (UTC) [ reply ] I agree that places like this shouldn\'t be erased from history, but we need reliable sources about them if we\'re going to host an article saying anything. All we have is a couple of statements that it was a point on a railroad map and there was a short-lived post office in the vicinity, neither of which cuts it for WP:N purposes. If anyone can find more, then of course we can the article and say more. WeirdNAnnoyed ( talk ) 01:11, 25 October 2023 (UTC) Thank you for expanding the article; this is exactly what we need! WeirdNAnnoyed ( talk ) 02:20, 25 October 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article\'s talk page or in a deletion review ). No further edits should be made to this page.'</li><li>"The Political Machine 2020: WP:BEFORE turned up no critic reviews, Metacritic also shows nothing except one review from a non-reliable source (New Game Network). This game fails WP:GNG . λ Negative MP1 21:09, 22 February 2024 (UTC) This discussion has been included in the list of Video games-related deletion discussions . λ Negative MP1 21:09, 22 February 2024 (UTC) GNG . Waxworker ( talk ) 21:42, 22 February 2024 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Politics and United States of America . WC Quidditch ☎ ✎ 02:42, 23 February 2024 (UTC) [ reply ] Per source found by Waxworker. It shows the article passes GNG. ᴢxᴄᴠʙɴᴍ ( ᴛ ) 06:44, 23 February 2024 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page."</li></ul> |
## Uses
### Direct Use for Inference
First install the SetFit library:
```bash
pip install setfit
```
Then you can load this model and run inference.
```python
from setfit import SetFitModel
# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("research-dump/bge-base-en-v1.5_wikipedia_r_masked_wikipedia_r_masked")
# Run inference
preds = model("Simon Sues : – Laundry Pizza 03 ( d c̄ ) 15:37, 12 April 2023 (UTC) This discussion has been included in the deletion sorting lists for the following topics: Science fiction and fantasy , Comics and animation , and Anime and manga . – Laundry Pizza 03 ( d c̄ ) 15:37, 12 April 2023 (UTC) [ reply ] issues not addressed for 8 years; Google only throws up listings and associated official and/or fan social media accounts, meaning notable neutral sources are unlikely to be found, at most to Tokyopop as seemingly only notable attached page. BoomboxTestarossa ( talk ) 16:15, 12 April 2023 (UTC) [ reply ] . It's ok to mass prod such unreferenced fancruft. -- Piotr Konieczny aka Prokonsul Piotrus | reply here 16:26, 12 April 2023 (UTC) [ reply ] Mass prod was my mistake, I forgot no-one could see that I'd checked to see if anything was salvageable beforehand and put through too many at once. I have been put gently right by some kind people, and hopefully will not have to do anything with that many pages at once again! =) BoomboxTestarossa ( talk ) 20:46, 12 April 2023 (UTC) [ reply ] . ZERO sources used, I'm not wading through every Google mention with a person named Simon that sues people. TNT this article. Oaktree b ( talk ) 20:00, 12 April 2023 (UTC) [ reply ] Strongly agree with what’s been said above. This is a clear . Go4thProsper ( talk ) 08:14, 15 April 2023 (UTC) [ reply ] - literally unsourced and difficult to find anything. Bearian ( talk ) 19:11, 18 April 2023 (UTC) GNG . Pharaoh of the Wizards ( talk ) 08:25, 19 April 2023 (UTC) [ reply ] The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review ). No further edits should be made to this page.")
```
<!--
### Downstream Use
*List how someone could finetune this model on their own dataset.*
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Set Metrics
| Training set | Min | Median | Max |
|:-------------|:----|:-------|:-----|
| Word count | 61 | 469.53 | 6182 |
| Label | Training Sample Count |
|:------|:----------------------|
| 0 | 542 |
| 1 | 163 |
| 2 | 57 |
| 3 | 65 |
| 4 | 127 |
| 5 | 13 |
| 6 | 24 |
| 7 | 9 |
### Training Hyperparameters
- batch_size: (8, 2)
- num_epochs: (5, 5)
- max_steps: -1
- sampling_strategy: oversampling
- num_iterations: 10
- body_learning_rate: (1e-05, 1e-05)
- head_learning_rate: 5e-05
- loss: CosineSimilarityLoss
- distance_metric: cosine_distance
- margin: 0.25
- end_to_end: True
- use_amp: True
- warmup_proportion: 0.1
- l2_weight: 0.01
- seed: 42
- eval_max_steps: -1
- load_best_model_at_end: False
### Training Results
| Epoch | Step | Training Loss | Validation Loss |
|:------:|:-----:|:-------------:|:---------------:|
| 0.0004 | 1 | 0.1964 | - |
| 0.2 | 500 | 0.2423 | 0.2325 |
| 0.4 | 1000 | 0.1902 | 0.2467 |
| 0.6 | 1500 | 0.115 | 0.2894 |
| 0.8 | 2000 | 0.0906 | 0.2921 |
| 1.0 | 2500 | 0.0604 | 0.3079 |
| 1.2 | 3000 | 0.0524 | 0.3013 |
| 1.4 | 3500 | 0.0427 | 0.3034 |
| 1.6 | 4000 | 0.0356 | 0.2897 |
| 1.8 | 4500 | 0.0302 | 0.3032 |
| 2.0 | 5000 | 0.0282 | 0.3159 |
| 2.2 | 5500 | 0.0183 | 0.3201 |
| 2.4 | 6000 | 0.0173 | 0.3163 |
| 2.6 | 6500 | 0.0128 | 0.3170 |
| 2.8 | 7000 | 0.0139 | 0.3038 |
| 3.0 | 7500 | 0.0135 | 0.3007 |
| 3.2 | 8000 | 0.01 | 0.3074 |
| 3.4 | 8500 | 0.0082 | 0.3086 |
| 3.6 | 9000 | 0.0071 | 0.3084 |
| 3.8 | 9500 | 0.007 | 0.2967 |
| 4.0 | 10000 | 0.0055 | 0.2932 |
| 4.2 | 10500 | 0.0046 | 0.2971 |
| 4.4 | 11000 | 0.0044 | 0.2901 |
| 4.6 | 11500 | 0.0046 | 0.2936 |
| 4.8 | 12000 | 0.004 | 0.2950 |
| 5.0 | 12500 | 0.0044 | 0.2955 |
### Framework Versions
- Python: 3.12.7
- SetFit: 1.1.1
- Sentence Transformers: 3.4.1
- Transformers: 4.48.2
- PyTorch: 2.6.0+cu124
- Datasets: 3.2.0
- Tokenizers: 0.21.0
## Citation
### BibTeX
```bibtex
@article{https://doi.org/10.48550/arxiv.2209.11055,
doi = {10.48550/ARXIV.2209.11055},
url = {https://arxiv.org/abs/2209.11055},
author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Efficient Few-Shot Learning Without Prompts},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
google/paligemma-3b-mix-224-keras | google | image-text-to-text | [
"keras-hub",
"image-text-to-text",
"license:gemma",
"region:us"
] | 1,719,427,626,000 | 2024-10-28T21:57:11 | 4 | 1 | ---
library_name: keras-hub
license: gemma
pipeline_tag: image-text-to-text
extra_gated_heading: Access PaliGemma on Hugging Face
extra_gated_prompt: To access PaliGemma on Hugging Face, you’re required to review
and agree to Google’s usage license. To do this, please ensure you’re logged-in
to Hugging Face and click below. Requests are processed immediately.
extra_gated_button_content: Acknowledge license
---
PaliGemma is a set of multi-modal large language models published by Google based on the Gemma model. Both a pre-trained and instruction tuned models are available. See the model card below for benchmarks, data sources, and intended use cases.
## Links
* [PaliGemma API Documentation](https://keras.io/api/keras_hub/models/pali_gemma/)
* [KerasHub Beginner Guide](https://keras.io/guides/keras_hub/getting_started/)
* [KerasHub Model Publishing Guide](https://keras.io/guides/keras_hub/upload/)
## Installation
Keras and KerasHub can be installed with:
```
pip install -U -q keras-hub
pip install -U -q keras>=3
```
Jax, TensorFlow, and Torch come preinstalled in Kaggle Notebooks. For instruction on installing them in another environment see the [Keras Getting Started](https://keras.io/getting_started/) page.
## Presets
The following model checkpoints are provided by the Keras team. Full code examples for each are available below.
| Preset name | Parameters | Description |
|-----------------------|------------|-------------------------------------------------------------|
| [**paligemma-3b-224-mix-keras**](https://huggingface.co/google/paligemma-3b-224-mix-keras) | 2.92B | image size 224, mix fine tuned, text sequence length is 256 |
| [paligemma-3b-448-mix-keras](https://huggingface.co/google/paligemma-3b-448-mix-keras) | 2.92B | image size 448, mix fine tuned, text sequence length is 512 |
| [paligemma-3b-224-keras](https://huggingface.co/google/paligemma-3b-224-keras) | 2.92B | image size 224, pre trained, text sequence length is 128 |
| [paligemma-3b-448-keras](https://huggingface.co/google/paligemma-3b-448-keras) | 2.92B | image size 448, pre trained, text sequence length is 512 |
| [paligemma-3b-896-keras](https://huggingface.co/google/paligemma-3b-896-keras) | 2.93B | image size 896, pre trained, text sequence length is 512 |
## Prompts
The PaliGemma `"mix"` models can handle a number of prompting structures out of the box. It is important to stick exactly to these prompts, including the newline. Lang can be a language code such as `"en"` or `"fr"`. Support for languages outside of English will vary depending on the prompt type.
* `"cap {lang}\n"`: very raw short caption (from WebLI-alt).
* `"caption {lang}\n"`: coco-like short captions.
* `"describe {lang}\n"`: somewhat longer more descriptive captions.
* `"ocr\n"`: optical character recognition.
* `"answer en {question}\n"`: question answering about the image contents.
* `"question {lang} {answer}\n"`: question generation for a given answer.
* `"detect {thing} ; {thing}\n"`: count objects in a scene.
Not `"mix"` presets should be fine-tuned for a specific task.
```
!pip install -U -q keras-hub
```
Pick a backend of your choice
```
import os
os.environ["KERAS_BACKEND"] = "jax"
```
Now we can load the PaliGemma "causal language model" from the Kaggle Models hub. A causal language model is just a LLM that is ready for generation, by training with a causal mask, and running generation a token at a time in a recurrent loop.
```
keras.config.set_floatx("bfloat16")
pali_gemma_lm = keras_hub.models.PaliGemmaCausalLM.from_preset(
"hf://google/paligemma-3b-224-mix-keras"
)
```
Function that reads an image from a given URL
```
def read_image(url):
contents = io.BytesIO(requests.get(url).content)
image = PIL.Image.open(contents)
image = np.array(image)
# Remove alpha channel if neccessary.
if image.shape[2] == 4:
image = image[:, :, :3]
return image
```
```
image_url = 'https://storage.googleapis.com/keras-cv/models/paligemma/cow_beach_1.png'
image = read_image(image_url)
```
Use `generate()` call with a single image and prompt. The text prompt
has to end with `\n`.
```
prompt = 'answer en where is the cow standing?\n'
output = pali_gemma_lm.generate(
inputs={
"images": image,
"prompts": prompt,
}
)
print(output)
```
Use `generate()` call with a batched images and prompts.
```
prompts = [
'answer en where is the cow standing?\n',
'answer en what color is the cow?\n',
'describe en\n',
'detect cow\n',
'segment cow\n',
]
images = [image, image, image, image, image]
outputs = pali_gemma_lm.generate(
inputs={
"images": images,
"prompts": prompts,
}
)
for output in outputs:
print(output)
```
There's a few other style of prompts this model can handle out of the box...
`cap {lang}\n`: very raw short caption (from WebLI-alt).
`caption {lang}\n`: nice, coco-like short captions.
`describe {lang}\n`: somewhat longer more descriptive captions.
`ocr\n`: optical character recognition.
`answer en {question}\n`: question answering about the image contents.
`question {lang} {answer}\n`: question generation for a given answer.
`detect {thing} ; {thing}\n`: count objects in a scene.
Call `fit()` on a single batch
```
import numpy as np
image = np.random.uniform(-1, 1, size=(224, 224, 3))
x = {
"images": [image, image],
"prompts": ["answer en Where is the cow standing?\n", "caption en\n"],
}
y = {
"responses": ["beach", "A brown cow standing on a beach next to the ocean."],
}
pali_gemma_lm = keras_hub.models.PaliGemmaCausalLM.from_preset("hf://google/paligemma-3b-224-mix-keras")
pali_gemma_lm.fit(x=x, y=y, batch_size=2)
``` | [
"QUESTION_ANSWERING"
] | Non_BioNLP |
Areeb123/En-Fr_Translation_Model | Areeb123 | translation | [
"transformers",
"tensorboard",
"safetensors",
"marian",
"text2text-generation",
"translation",
"generated_from_trainer",
"en",
"fr",
"dataset:kde4",
"base_model:Helsinki-NLP/opus-mt-en-fr",
"base_model:finetune:Helsinki-NLP/opus-mt-en-fr",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,701,002,848,000 | 2023-11-27T12:58:18 | 38 | 0 | ---
base_model: Helsinki-NLP/opus-mt-en-fr
datasets:
- kde4
language:
- en
- fr
license: apache-2.0
metrics:
- bleu
tags:
- translation
- generated_from_trainer
model-index:
- name: En-Fr_Translation_Model
results:
- task:
type: text2text-generation
name: Sequence-to-sequence Language Modeling
dataset:
name: kde4
type: kde4
config: en-fr
split: train
args: en-fr
metrics:
- type: bleu
value: 52.78125912187245
name: Bleu
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# En-Fr_Translation_Model
This model is a fine-tuned version of [Helsinki-NLP/opus-mt-en-fr](https://huggingface.co/Helsinki-NLP/opus-mt-en-fr) on the kde4 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8567
- Bleu: 52.7813
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
### Training results
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0 | [
"TRANSLATION"
] | Non_BioNLP |
Bijayab/a100_80_nepberta | Bijayab | text2text-generation | [
"transformers",
"safetensors",
"bart",
"text2text-generation",
"summary",
"nepali",
"BART",
"NLP",
"ne",
"dataset:csebuetnlp/xlsum",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,711,130,672,000 | 2024-06-15T15:09:00 | 28 | 0 | ---
datasets:
- csebuetnlp/xlsum
language:
- ne
library_name: transformers
license: mit
metrics:
- rouge
pipeline_tag: text2text-generation
tags:
- summary
- nepali
- BART
- NLP
---
from transformers import pipeline
input_text = """सांसदको लोगो छोडेर सिलाम साक्मामा मात्र भेटिएपछि उनलाई जिज्ञाशा राखियो ।उनले किराती परम्परामा सिलाम साक्माको महत्त्वबारे उल्लेख गर्दै 'कान्तिपुर' को स्मरणका लागि 'सिलाम साक्मा' लगाएको तस्बिर खिच्न आग्रह गरे । स्मरणका लागि भन्दै खिचाएको त्यही तस्बिर जस्तै उनले राजनीतिमा खेलेको भूमिका र योगदान पनि अब इतिहासको गर्भमा पुगेको छ ।
अब शान्त, शालीन र भद्र राजनीतिज्ञका रूपमा धेरैले चिन्ने तिनै सादगी सुवासचन्द्र नेम्वाङको भौतिक शरीर अब रहेन । सोमबार मध्यराति उनको निधन भयो । उनको निधनप्रति धेरै राजनीतिज्ञ, सामाजिक क्षेत्रका व्यक्ति लगायतले गहिरो शोक व्यक्त गरेका छन् । सक्रिय राजनीतिक गतिविधिमा संलग्न रहँदा रहँदै भएको उनको निधनले एमालेजन मात्र होइन धेरैलाई स्तब्ध बनाएको छ । नेम्वाङप्रतिका श्रद्धाञ्जलीका शब्द र तस्बिरले सामाजिक सञ्जालहरू पनि शोकमय भएका छन् ।
इलामको सुन्तलाबारीमा २००९ सालमा जन्मिएका नेम्वाङको ७० वर्षे जीवनका पछिल्ला ५० वर्ष भने सक्रिय राजनीतिमा बिते । कानुनी पृष्टभूमिसहित पार्टी राजनीतिमा सक्रिय रहेका नेम्वाङ संविधानसभाबाट संविधान जारी गर्ने हरेक घटनाक्रममा एक प्रत्यक्ष साक्षी र निर्णयकर्ता हुन् । संविधानसभा, सभामुख र संसद् भन्नेबित्तिकै त्यसको पर्यायवाचीका रूपमा नाम लिइने व्यक्ति थिए, नेम्वाङ । दुइटै संविधानसभाका अध्यक्ष नेम्वाङले संविधानसभाबाट संविधान जारी गर्ने क्रममा खेलेको भूमिका स्मरणीय छ ।"""
summarizer = pipeline("summarization", model="Bijayab/a100_80_nepberta")
results = summarizer(input_text, max_length=130, min_length=30, do_sample=False)
print(results[0]['summary_text']) | [
"SUMMARIZATION"
] | Non_BioNLP |
Helsinki-NLP/opus-mt-fr-sk | Helsinki-NLP | translation | [
"transformers",
"pytorch",
"tf",
"marian",
"text2text-generation",
"translation",
"fr",
"sk",
"license:apache-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,646,263,744,000 | 2023-08-16T11:37:13 | 41 | 0 | ---
license: apache-2.0
tags:
- translation
---
### opus-mt-fr-sk
* source languages: fr
* target languages: sk
* OPUS readme: [fr-sk](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/fr-sk/README.md)
* dataset: opus
* model: transformer-align
* pre-processing: normalization + SentencePiece
* download original weights: [opus-2020-01-16.zip](https://object.pouta.csc.fi/OPUS-MT-models/fr-sk/opus-2020-01-16.zip)
* test set translations: [opus-2020-01-16.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/fr-sk/opus-2020-01-16.test.txt)
* test set scores: [opus-2020-01-16.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/fr-sk/opus-2020-01-16.eval.txt)
## Benchmarks
| testset | BLEU | chr-F |
|-----------------------|-------|-------|
| JW300.fr.sk | 24.9 | 0.456 |
| [
"TRANSLATION"
] | Non_BioNLP |
zyxzyx/autotrain-sum-1042335811 | zyxzyx | text2text-generation | [
"transformers",
"pytorch",
"mt5",
"text2text-generation",
"autotrain",
"zh",
"dataset:zyxzyx/autotrain-data-sum",
"co2_eq_emissions",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,656,293,128,000 | 2022-06-27T05:15:17 | 96 | 0 | ---
datasets:
- zyxzyx/autotrain-data-sum
language: zh
tags:
- a
- u
- t
- o
- r
- i
- n
widget:
- text: I love AutoTrain 🤗
co2_eq_emissions: 426.15271368095927
---
# Model Trained Using AutoTrain
- Problem type: Summarization
- Model ID: 1042335811
- CO2 Emissions (in grams): 426.15271368095927
## Validation Metrics
- Loss: 1.7748287916183472
- Rouge1: 0.536
- Rouge2: 0.0
- RougeL: 0.536
- RougeLsum: 0.536
- Gen Len: 10.9089
## Usage
You can use cURL to access this model:
```
$ curl -X POST -H "Authorization: Bearer YOUR_HUGGINGFACE_API_KEY" -H "Content-Type: application/json" -d '{"inputs": "I love AutoTrain"}' https://api-inference.huggingface.co/zyxzyx/autotrain-sum-1042335811
``` | [
"SUMMARIZATION"
] | Non_BioNLP |
seongil-dn/bge-m3-kor-retrieval-451949-bs64-finance-50 | seongil-dn | sentence-similarity | [
"sentence-transformers",
"safetensors",
"xlm-roberta",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:451949",
"loss:CachedMultipleNegativesRankingLoss",
"arxiv:1908.10084",
"arxiv:2101.06983",
"base_model:BAAI/bge-m3",
"base_model:finetune:BAAI/bge-m3",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,734,167,762,000 | 2024-12-14T09:17:18 | 82 | 0 | ---
base_model: BAAI/bge-m3
library_name: sentence-transformers
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:451949
- loss:CachedMultipleNegativesRankingLoss
widget:
- source_sentence: 사설묘지의 관리방법에 대한 27.9퍼센트의 견해에 근거하면 모든 묘지는 어떻게 관리해야 해?
sentences:
- '(10) 공원묘지의 활성화 방안
□ 문제점
공원묘지의 재정비, 활성화는 시급하고 적절한 방안이나 법률에는 규정이 없다. □ 개선방안
첫째 공원묘지 중 무연고 분묘는 절차를 거쳐 화장을 하고 그 골분을 묘지 내 봉안소로 안치하고 분묘자리는 자연장지 또는 다른 장지로 활용한다.
둘째 장사법이 정한 분묘기간보다 이전인 60년 이상 된 묘지에 대해서는 적극적으로 개장을 유도한다. 다만, 연고자를 찾지 못하는 경우에는 무연고
자에 준하여 처리할 수 있는 규정을 두어 처리한다. (11) 사설묘지의 조성기준의 포괄적 위임
□ 문제점
법 제14조(사설묘지의 설치 등) 제6항은 “사설묘지의 설치면적, 분묘의 형태, 설치장소, 그 밖의 설치기준 등에 관하여 필요한 사항은 대통령령으로
정한다.” 라고 포괄적으로 위임을 하고 있다. 그러나 대통령령에 의한 [별표]에서 종중이나 문중의 묘지와 봉안묘의 설치 수와 특히 종교단체의
봉안묘의 설치수가 각 1개소로 규정하는 것은 대단히 중요한 내용으로 법률에 직접 규정해야 한다는 의견이 있다. □ 개선방안
사설묘지의 조성기준 중 사설묘지, 사설봉안당 등 중요한 사항은 법률로 정하는 것이 합리적이다. * 의견수렴 결과
(2-1-1) 묘지의 설치 제한(법 제17조)
장사법은 ‘묘지’의 설치와 지한은 ‘화장장과 봉안당, 자연장’과 동일하게 규정하고 있어 설치를 어렵게 한다는 의견이 있는데, 이에 대한 응답으로는
묘지, 화장장, 봉안당, 자연장은 성격이 다르므로 구분하여 규정해야 한다는 의견이 전체의 84.1%로 현행대로 좋다는 의견 15.9%보다 월등히
높게 나타났다. (2-2-1) 개인소유 토지의 묘지조성
개인이 소유한 토지에 ‘사설묘지’를 조성하고 임의로 매장하는 것이 적절한가에 대한 응답으로는 사유지라도 임의로 설치하는 것은 제한해야 한다는
의견이 전체의 68.8%로 가장 높게 나타났다. (2-3-1) 묘지의 성격 구분(법 제14조)-중복
법 제14조에 규정한 종중, 문중묘지, 법인묘지 규정에서 종중과 문중, 법인의 성격을 구체적으로 구분해야 한다는 의견에 대해서는 더 구체적으로
규정해야 한다는 의견이 전체의 63.1%로 가장 높게 나타났다. (2-4-1) 묘지와 상석 등의 규모, 수
추모문화와 관습은 지역, 문화, 관습, 종교에 따라 차이가 있는바, 묘지와 상석 등의 규모, 수를 법률로 정하는 것은 무리라는 의견이 있는데,
이에 대한 응답으로는 법률에 구체적으로 규정해야 한다는 의견이 전체의 62.1%로 가장 높게 나타났다. (2-5-1) 분묘의 설치기간(법 제19조)
장사법은 설치기간을 15년에서 다시 15년씩 3회 연장하여 총 60년을 사용할 수 있게 규정하고 있는데, 이에 대해서는 기간을 단축해야 한다는
의견이 전체의 40.6%로 가장 높게 나타났고, 그 뒤를 적절한 규정이라는 의견과 제한자체가 필요없다는 의견이 함께 21.9%를 차지했다.
(2-5-2) 사설묘지의 관리방법
사설묘지의 관리방법이 모호하여 시행가능성이 없다는 의견이 있는데, 이에 대해서는 묘지 설치 및 이전 등의 신고를 강제화하고 위반 시 벌칙을
강화해야 한다는 의견이 전체의 39.3%로 가장 높게 나타났으며, 그 다음으로는 모든 묘지를 전산화하여 관리해야 한다는 의견이 27.9%로
두 번째로 높게 나타났다.'
- 본고는 정부를 단일한 의사결정주체로 보기보다는 유인체계에 민감한 개인들로 구성된 조직이라는 인식 하에 정부실패를 해결하기 위한 방안으로 유인체계
설계의 중요성을 짚어보고 있다. 기존 경제학에서는 시장실패 문제에 대해 시장참여자의 유인체계가 강조되었는데, 감독실패와 같은 정부실패 문제에
대해서도 규제·감독권자의 유인체계를 중심으로 해결방안을 모색할 수 있을 것이다. 여기서는 주인-대리인(principal-agent) 모형의
관점에서 규제·감독권자의 의사결정이 감독권한을 위임한 국민(납세자·유권자)의 이익에 배치될 수 있음을 설명하고 있다. 국민과 규제·감독권자
사이에는 정보비대칭이 있으며, 규제·감독권자와 금융기관 사이에도 정보비대칭이 있다. 여기서 감독권자는 금융안정 등과 같이 국민의 이익에 부합하는
목적을 달성하기 위해 금융기관을 감독할 책무를 부여받고 있다. 한편, 개별 국민은 정보비대칭에 따른 정보 부족과 감독업무의 성과와 연계된 이익이
미미함에 따라 감독권자의 책무 이행을 감시할 유인이 부족하다. 이상의 환경 하에서 사적 이익(private interest)을 추구하는 감독권자는
규제 포획(regulatory capture)에 취약할 수 있다. 다시 말해 상기의 환경 하에서는 산업, 정치 등으로부터의 영향력 행사가 감독권자의
사익에 부합할 경우에 공익에서 벗어난 의사결정을 초래할 수 있다. 이상과 같이 왜곡된 유인체계로 인한 감독지배구조의 실패라는 문제를 해결하기
위해서는 감독권자의 유인체계를 조정하는 방안을 모색할 수 있을 것이다.
- '4. 개선과제
(1) 화장시설 운영의 개선과제
첫째, 지역주민들의 강력한 반대 속에 대규모 화장시설을 확충하는 방식을 지양하고, 환경 친화적인 소규모 추모공원의 형태로 화장시설을 늘려나가는
방식이 필요하다. 둘째, 현행 화장시설의 공공성을 고려하여 기타 사회복지시설에 포함시키고, 관련 비용의 감액이 가능하도록 입법화하는 방안도
모색할 필요성이 있다. 셋째, 장례업자로 하여금, 화장하는 시신의 경우는 유가족들에게 매장용이 아닌 화장 전용 관을 판매하도록 독려해야 할
것이다. (2) 봉안시설 운영의 개선과제
첫째, 봉안시설이 무분별하게 증가하지 못하도록 설치 및 운영을 현행 신고제에서 허가제로 변경하는 방안도 고려해 볼 수 있다. 이는 봉안시설의
환경훼손 문제를 해결하고 시설운영적자로 인한 빈번한 폐지 등을 방지하기 위해서 필요한 조치로 사료된다. 둘째, 봉안시설이 향후 관리되지 않고
방치되어 무연고 골분의 발생을 야기시킬 수 있는 만큼, 봉안시설을 사자(死者)에 대한 반영구적 추모시설로 활용할 방안 마련이 필요하다. (3)
자연장지 운영의 개선과제
첫째, 기존의 분묘 설치기간을 단축할 수 있도록 기준일을 기존 분묘 설치일로 변경하는 등의 제도개선을 통해 분묘의 개장을 독려하고 자연장지를
확대해 나가는 정책적 접근이 필요하다. 둘째, 자연장지의 조성에 있어서 걸림돌이 될 수 있는 규정들, 예컨대 기존의 묘지지역에는 다양한 유형의
묘지를 운영해 나갈 수 있도록 절차를 단순화하려는 노력이 필요하다. 이러한 개선과제들은 장사시설을 혐오하는 사고방식을 탈피하고, 진정한 추모의식과
문화를 계승해 나가려는 국민들의 적극적인 노력이 밑받침될 때 비로소 성공할 수 있을 것이다.'
- source_sentence: 중소기업의 수출을 정책적으로 지원하기 위해 어떻게 조사를 진행했어?
sentences:
- '4. 시사점
2000~2008년과 2009~2012년의 두 기간을 비교할 때, 중소ㆍ중견기업과 대기업의 수출 증가율 차이는 4.5%만큼 확대되었는데, 이는
대기업의 높은 수출 증가율에 기인하는 바가 크다. 한편 2009년 이후 대기업은 수출비중 및 수출 증가율이 증가한 것에 반하여 중소기업의 수출
비중과 매년 수출증가율은 감소했다. 2012년에 중소기업 수출증가율이 1.1%에 불과함을 고려하면 영세업체가 대다수인 중소기업의 수출 증대가
난관에 봉착해 있으며, 이의 돌파를 위한 다양한 정책적 지원이 필요함을 알 수 있다. 효율적인 정책적 지원을 위해 수출 중소기업의 애로 및
필요가 어떻게 변천되었는지 검토한 결과, 외환위기 시기인 1998년의 수출 중소기업들은 수출시작 단계인 업체와 수출 중인 업체 모두 자금지원의
필요성을 순위 높게 매겼으며 수출시작 시기인 중소기업들은 바이어찾기 등 수출 성사에 직접적으로 연관된 정보에 대한 필요가 컸다. 2012년
및 2014년의 수출 중소기업들은 자금 지원보다는 마케팅 지원 및 정보제공 확대를 공통적으로 최우선 순위로 요구하였다. 그러나 1998년에
자금 지원의 필요가 가장 두드러졌던 것은 외환위기라는 특수한 시기에 따른 일시적인 현상으로 볼 수 있다. 한편 해외바이어정보 및 무역 동향ㆍ해외시장정보
제공에 해당하는 ‘정보제공 및 판매인프라 구축’ 항목과 무역사절단 파견ㆍ해외전시회 참가ㆍ시장개척단 파견 등에 해당하는 ‘해외 마케팅’ 항목에
대한 중소기업의 필요는 항상 높은 순위를 기록하였다. 최근 이에 관한 필요성이 더욱 두드러지므로 관련 민ㆍ관 기관들의 원활한 업무연계가 더욱
요구된다. 특히 산업통상자원부는 바이어 정보 DB 및 국내 업체 DB를 보유한 코트라, 중소기업진흥공사, 무역협회 등 연관 기관들이 기관 각각의
이해관계를 넘어서서 수요자인 중소기업이 수출현장에서 높은 협상력을 보유하도록 장기적으로 DB를 공유ㆍ통합할 수 있는 효율적이고 합리적인 방안을
강구해야 할 것이다. 또한 해외 마케팅 관련 예산을 증액하고 신흥시장 수출 증가를 위하여 국격제고 및 한류와 CSR 이용 등 다양한 수출마케팅
전략 추진을 지속ㆍ확대할 필요가 있다.'
- '보험업법 시행령 및 감독규정 일부개정령(안) 입법예고
1. 개정이유
보험업법 규제입증정비위원회 등을 통해 정비 필요성이 입증된 기존 규제를 정비하는 한편 손해사정 제도 정비를 통해 보험소비자의 권익을 제고하기
위함 2. 주요 내용
가. 보험회사의 중요사항 설명의무 확대
보험회사가 상법에 따른 보험금청구권의 소멸시효, 손해사정사 선임에 관한 동의기준 등을 소비자에게 안내하도록 의무화
나. 보험회사의 선불전자지급업무 겸영 허용
보험회사가 헬스케어 서비스 운영을 위해 필요한 범위에서 선불전자지급업무를 겸영업무로 영위할 수 있도록 허용
다. 손해사정업자의 책임성 강화
손해사정 업무의 공정성·책임성 강화를 위해 손해사정협회가 표준 업무기준을 마련하여 손해사정업자에 권고토록 하고, 대형 손해사정업자(100인
이상)에 대해서는 금융위·원이 정하는 세부 업무기준·요건을 의무적으로 갖추도록 규정
라. 보험업 인허가 심사중단제도 개선
보험업 인허가 관련 심사지연을 방지하고 신청인의 예측가능성을 제고하기 위해 중대성·명백성 등 기본원칙에 따라 중단요건을 세분화·구체화하고,
매 6개월마다 심사재개여부를 검토하도록 규정'
- "하반기 수출 단기 지원 프로그램으로 , Cheer-up! □ 지식경제부는 7.25(수) 홍석우 장관 주재로 어려워지는 하반기 중소기업\
\ 수출여건에 선제적으로 대응하기 위해 ‘중소기업 수출 확대 지원회의 ’ 를 갖고 , ｢중소기업 수출확대 단기지원방안 ｣ 을 발표함 \nㅇ\
\ 회의에는 수출 중소기업과 KOTRA, K-Sure등 유관기관에서 참석하여 하반기 악화되는 대외 수출여건 대응방안에 대해 논의함\n□ 최근\
\ 대외여건 악화로 2012년 상반기 우리 무역은 수출입 증가율이 둔화되고 무역흑자가 2/3 수준으로 축소되는 등 불안한 모습\nㅇ 상반기에\
\ 이어 하반기에도 그리스, 스페인 등 EU 재정위기로부터의 불안요인이 지속되는 가운데, 수출 점유율이 높은 미국과 중국의 경기 회복 지연과\
\ 최대 수출품목인 선박 수출 부진이 우려되는 상황"
- source_sentence: 어떻게 입법부는 통일비용에 대한 정치사회학적인 접근을 도모할 수 있어?
sentences:
- '3. 정책적 측면
우선 통일비용에 대한 정치사회적 접근을 이루기 위해서는 국회가 통일정책과 대북정책의 상관성을 명확히 재정립하는 것이 필요하다. 역대 정부는
항상 통일정책이 대북정책과 밀접하게 연관되어 있다고 표방해왔다. 즉 평화통일이라는 헌법적 명령에 따라 포괄적인 범주를 가진 통일정책을 추진하며,
이틀 속에서 북한을 대상으로 한 대북정책이 입안·추진된다는 것이다. 이의 당위성은 의심의 여지가 없지만, 이제까지 실제로 추진된 정책도 과연
그러했었는지는 의문이다(김학성, 2012). 더욱이 통일 문제는 겉으로 보기에는 민족 문제이고 국제 문제이지만 DNA상으로는 정치 문제이기
때문에, 통일 문제는 신보수주의의 사상적 기원이 되고 있는 슈미트(Karl Schmidt)의 말(카를 슈미트, 2012)처럼, 국내적으로는
승자 독식의 정치판 세계에서 ‘적과 동지의 문제’가 되어 극과 극의 대결이 치닫게 되는 경우가 대부분이다(정세현, 2013).'
- '전통주 등의 산업진흥에 관한 법률 시행령 일부개정령안 입법예고
1. 개정이유
전통주 등 관련 단체의 자조금 적립 지원 등의 내용으로 전통주 등의 산업진흥에 관한 법률이 개정(법률 제16788호, 2019. 12. 10.
공포, 2020. 6. 11. 시행)됨에 따라, 자조금의 조성방법, 보조금의 지급기준 등 법률에서 위임된 사항과 그 시행을 위하여 필요한 사항을
정하려는 것임 2. 주요내용
가. 자조금의 조성방법 및 용도(안 제4조의2)
1) (제·개정 주요내용) 자조금 조성 단체는 그 구성원이 자율적으로 납입하는 금액으로 자조금을 조성하고, 전통주 등의 홍보, 품질향상, 판로확대
등을 위해 사용하도록 정함 2) (제·개정 사유) 자조금 사업을 통해 전통주 등 생산업체의 경쟁력이 제고될 것으로 기대됨
나. 보조금 지급 기준(안 제4조의3)
1) (제·개정 주요내용) 보조금을 지급 받으려는 자조금 조성 단체의 요건, 보조금의 지급한도, 관련 자료제출 요청 권한 등을 정함 2) (제·개정
사유) 자조금의 효과적인 운영을 위해 보조금 지급기준을 정함'
- 여야는 통일비용에 대한 주요 쟁점 사안에 대해서는 사전에 협력을 모색해야 한다. 여권은 정부의 입장을 무조건 지지하기보다는 국민통합이라는 차원에서
심사숙고해야 하고, 야권은 정부의 입장을 그대로 수용하기 어렵다고 하더라도 반대를 위한 반대를 지양하고 대안을 제시해야 한다. 따라서 여야는
여야 간사 회의 등을 통해 통일비용에 대한 주요 사안에 대해서 사전 조율하는 노력이 필요하다. 통일비용에 대한 초당적인 협력이 구현되기 위해
국회차원의 제도적인 조정·협의 기구(분과위원회 개최 등)도 필요하다. 정치권은 다수 국민이 무엇을 생각하는지를 우선 고려해서 정책방향을 수립해야
한다. 초당적인 협력을 바탕으로 하는 통일정책은 정권이 교체된다고 하더라도 지속될 수 있다. 통일정책에 대한 정치권의 초당적인 협력을 바탕으로
통일정책에 대한 국민적 화합을 모색하는 것이 필요하다. 북한과의 군사적 대치 상황에서 안보는 그 무엇과 바꿀 수 없을 정도로 매우 중요한 사안이다.
그렇지만 안보 위주의 통일정책은 남북관계의 발전을 요구하는 다수 국민의 수요를 충족시키기에는 부족함이 없지 않다. 통일정책은 남북관계의 발전과
안보를 동시에 고려하면서 적절한 합일점을 찾아야 한다. 정부의 성향에 따라 통일정책이 강경과 온건을 선회하는 현상을 반복해서는 국민적 합의를
이끌어낼 수 없다.
- source_sentence: 경북 동해안지역 비은행의 2010년 기준 수신 증가율이 어떻게 돼?
sentences:
- □ 예금은행과 비은행금융기관 모두 금융위기의 영향으로 증가세가 일시 낮아졌으나 2008년 이후 증가세를 회복한 가운데 비은행금융기관의 증가세
확대가 현저<br>ㅇ 시장이자율 하락에 따른 상대적 고금리 매력 부각, 2009년중 비과세한도 상향 조정 등에 따라 비은행금융기관의 수신이
상대적으로 호조를 보인 점은 전국적인 현상이나 2010년 들어 전국의 예금은행 수신이 빠르게 증가하고 비은행금융기관 수신 증가세가 상당폭 둔화된
것과는 달리 경북동해안은 비은행 금융기관이 비교적 높은 증가세를 유지하고 예금은행 증가세는 상대적으로 완만<br>― 이는 2010년 들어 상호저축은행의
부실 우려가 확산되면서 자금이 예금은행으로 이동함에 따라 전국 비은행금융기관의 수신 증가세가 빠르게 위축된 반면 경북동해안은 주민들이 이미
2007년 경북상호저축은행 사태를 겪<br>은 데다 상호저축은행의 비중도 낮아 크게 동요되지 않았던 데 주로 기인<br>ㅇ 다만 2011년
들어 타 지역 부실저축은행 영업정지 사태, 지역내 일부 비은행금융기관의 횡령 사고 등으로 비은행금융기관에 대한 신뢰가 하락하면서 역내 비은행금융기관의
수신 증가세도 주춤
- '□ 경북동해안지역 예금은행의 가계대출 증가율은 전국 예금은행 뿐만 아니라 전국 비은행금융기관의 증가율도 상회하고 있음
ㅇ 전국 예금은행과 비은행금융기관의 경우 가계대출 증가세가 2013.4월과 5월에 들어서야 회복되었으나 경북동해안지역은 2009년말부터 동
대출 증가세가 꾸준히 확대되었음
ㅇ 지역내 예금은행의 가계대출 증가세는 2012년초부터 주택담보대출에 의해 견인되고 있는 모습임
― 경북동해안지역 예금은행의 가계대출액은 3조 3,832억원으로 주택담보대출액은 이중 54.5%인 1조 8,436억원임(2013.9월말 잔액기준)'
- '2. 일수벌금제와 책임주의
일수벌금제 도입을 반대하는 입장에서의 가장 근본적인 근거는 일수벌금제는 동일한 (불법)행위에 대하여 동일한 책임을 지는 것이 아니기 때문에
이것이 형법상 책임주의 원칙에 반한다는 것이다. 즉, 행위 책임에 있어서 행위자 책임 요소인 경제적 상황을 고려하는 것은 타당하지 않다는 것이다.
이러한 입장에서는 동일한 범죄에 대하여는 행위자의 상황과 관계없이 동일한 벌금형을 부과하는 것이 평등하다고 본다. 반면 일수벌금제 도입에 찬성하는
입장에서는 일수벌금형제도는 형벌에 있어서 동일한 형벌효과를 줄 수 있는 제재를 부과하는 것이 형벌상의 실질적 평등을 실현한다는 측면에서 책임주의에
반하지 않는다고 주장한다. 동일한 형벌효과(형벌효과의 형평성)를 주기 위해 행위 책임의 범위 안에서 행위자 책임을 고려하는 것은 책임주의에
반하지 않는다는 것이다.'
- source_sentence: 빅데이터 활용의 증가로 인해 개인정보를 보호하기 위한 조치로 도입하려고 하는 건 뭐야?
sentences:
- 후속 질문으로 ‘현충시설 대체 명칭 선호도’를 공통적으로 물었으며, 응답자 편의를 위하여 설문지 내에 현충시설의 정의를 제시하여 이해를 도왔다.
민주화운동 관련 기념시설 포함 여부와 유사하게 명칭 선호도에서도 일반인과 공무원의 시각 차이가 드러났다. 일반인은 ‘현충선양시설’(22%),
‘현충민주시설’(16.5%), ‘현충사적’(15.5%)을 순서대로 선택하였지만, 공무원은 ‘보훈선양시설’(36.1%), ‘현충시설’(27.9%),
‘현충선양시설’(16.4%) 순으로 선호도가 나타났다. 일반인 1∼3순위 모두에 ‘현충’이 들어가 기존 현충시설 명칭의 연장선으로 인지하는
듯 보였으나, 이와 달리 공무원은 ‘보훈’, ‘선양’ 등 부서별이나 정책에서 사용되는 명칭을 선호하는 것으로 드러났다(<표 7> 참고).
- 'Ⅰ. 서론
□ 최근 빅데이터(Big Data) 분석 등을 통해 다양한 사회적 유용성과 가치를 가지는 정보 산출(산업)에 관심이 높아지면서, 개인정보 보호규제가
장애요인으로 등장하고 있다는 평가로 인하여, 비식별화(de-identification) 또는 익명화(anonymization)를 통한 개인정보
활용의 필요성이 논의되고 있음 ○ 국내외 개인정보 보호법제들은 기본적으로 ‘개인 식별 가능성’을 판단 기준으로 하여 개인정보의 개념적 범위를
설정하여 이를 보호하는 체계를 가지고 있으며, 일반적으로 이러한 개인 식별 가능성을 제약 또는 제거하는 조치를 ‘비식별화’ 또는 ‘익명화’라고
표현함
○ 우리나라뿐만 아니라, EU, 미국, 일본 등 주요 국가들도 빅데이터 등의 산업적 분석 및 활용 과정에서 발생하고 있는 문제점에 착안하여
개인정보 보호 법제 개선에 관한 논의를 진지하게 진행하고 있음 □ 국내외적으로 개인정보 ‘비식별화’ 또는 ‘익명화’, 그리고 그 법적 효과와
관련해서는 아직까지 확고한 개념정의 등이 존재하지 않는 상황이라고 할 수 있어, 관련 개념의 법적 활용에 있어 상당한 논란이 제기되고 있음
○ ‘비식별화’는 특정 정보로부터 ‘개인 식별 가능성’을 제거하는 조치 및 과정을 의미하며, 종종 이러한 용어는 ‘익명화’와 동일한 의미로
사용됨
- 비식별화라는 용어는 미국 등의 국가에서 주로 활용되고 있으며, 익명화 또는 익명가공이라는 용어는 EU 및 일본 등지에서 활용되고 있는 것으로
파악되지만, 대체적으로 두 용어가 혼용되고 있는 양상임'
- □ 지난 2014년 1월 발생한 카드회사의 개인정보 유출사고 이후 개인정보의 보호 강화 및 정보유출 재발방지를 위한 제도개선 방안이 논의되어
「금융분야 개인정보 유출 재발방지 종합대책」이 마련되기도 하였음 ○ 이와 함께 2015년 7월 「개인정보보호법」이 개정되어 징벌적손해배상제도가
도입되고, 불법적인 개인정보 유통 등에 대한 제재수준이 강화되었음 □ 하지만 지난 카드사태의 개인정보 유출 사례와 같이 한번 유출된 정보는
회수가 불가능하고 사후적인 제재를 강화하는 것만으로는 금융소비자의 피해를 방지하는데 한계가 있을 수밖에 없음 ○ 이에 따라 정보의 수집, 제공,
유통, 관리 전반에서 있어서 금융회사의 개인정보에 대한 기술적‧관리적 보호체계의 실질적인 강화를 통한 금융사고 예방이 보다 중요하다고 할 수
있음 □ 한편 우리나라는 해외와 비교해보면 빅데이터 활용이 초기단계에 해당하지만, 인터넷전문은행 도입과 관련하여 빅데이터 활용에 대한 금융회사의
관심이 증가하고 있음 ○ 현재 빅데이터를 활성화하기 위해 「신용정보의 이용 및 보호에 관한 법률」 개정을 통해 개인신용정보의 개념을 보다 명확하게
정의하여 금융회사가 활용할 수 있는 비식별정보를 구분하는 방안이 추진되고 있음 □ 하지만 개인신용정보 중에서 식별화정보와 비식별화 정보를 명확히
구분하는 것이 현실적으로 쉽지 않으므로 빅데이터의 활성화를 위해 개인정보의 균형있는 보호와 활용이 필요한 시점임 ○ 특히 개인신용정보에서 개인을
식별할 수 없는 비식별정보를 구분하는 경우에도 동 정보의 이용 목적을 제한하고, 비식별정보의 재식별 방지를 위한 안전성 확보 방안 등을 마련할
필요가 있음
---
# SentenceTransformer based on BAAI/bge-m3
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3) <!-- at revision 5617a9f61b028005a4858fdac845db406aefb181 -->
- **Maximum Sequence Length:** 1024 tokens
- **Output Dimensionality:** 1024 tokens
- **Similarity Function:** Cosine Similarity
<!-- - **Training Dataset:** Unknown -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 1024, 'do_lower_case': False}) with Transformer model: XLMRobertaModel
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("seongil-dn/bge-m3-kor-retrieval-451949-bs64-finance-50")
# Run inference
sentences = [
'빅데이터 활용의 증가로 인해 개인정보를 보호하기 위한 조치로 도입하려고 하는 건 뭐야?',
'Ⅰ. 서론\n□ 최근 빅데이터(Big Data) 분석 등을 통해 다양한 사회적 유용성과 가치를 가지는 정보 산출(산업)에 관심이 높아지면서, 개인정보 보호규제가 장애요인으로 등장하고 있다는 평가로 인하여, 비식별화(de-identification) 또는 익명화(anonymization)를 통한 개인정보 활용의 필요성이 논의되고 있음 ○ 국내외 개인정보 보호법제들은 기본적으로 ‘개인 식별 가능성’을 판단 기준으로 하여 개인정보의 개념적 범위를 설정하여 이를 보호하는 체계를 가지고 있으며, 일반적으로 이러한 개인 식별 가능성을 제약 또는 제거하는 조치를 ‘비식별화’ 또는 ‘익명화’라고 표현함\n○ 우리나라뿐만 아니라, EU, 미국, 일본 등 주요 국가들도 빅데이터 등의 산업적 분석 및 활용 과정에서 발생하고 있는 문제점에 착안하여 개인정보 보호 법제 개선에 관한 논의를 진지하게 진행하고 있음 □ 국내외적으로 개인정보 ‘비식별화’ 또는 ‘익명화’, 그리고 그 법적 효과와 관련해서는 아직까지 확고한 개념정의 등이 존재하지 않는 상황이라고 할 수 있어, 관련 개념의 법적 활용에 있어 상당한 논란이 제기되고 있음 ○ ‘비식별화’는 특정 정보로부터 ‘개인 식별 가능성’을 제거하는 조치 및 과정을 의미하며, 종종 이러한 용어는 ‘익명화’와 동일한 의미로 사용됨\n- 비식별화라는 용어는 미국 등의 국가에서 주로 활용되고 있으며, 익명화 또는 익명가공이라는 용어는 EU 및 일본 등지에서 활용되고 있는 것으로 파악되지만, 대체적으로 두 용어가 혼용되고 있는 양상임',
'□ 지난 2014년 1월 발생한 카드회사의 개인정보 유출사고 이후 개인정보의 보호 강화 및 정보유출 재발방지를 위한 제도개선 방안이 논의되어 「금융분야 개인정보 유출 재발방지 종합대책」이 마련되기도 하였음 ○ 이와 함께 2015년 7월 「개인정보보호법」이 개정되어 징벌적손해배상제도가 도입되고, 불법적인 개인정보 유통 등에 대한 제재수준이 강화되었음 □ 하지만 지난 카드사태의 개인정보 유출 사례와 같이 한번 유출된 정보는 회수가 불가능하고 사후적인 제재를 강화하는 것만으로는 금융소비자의 피해를 방지하는데 한계가 있을 수밖에 없음 ○ 이에 따라 정보의 수집, 제공, 유통, 관리 전반에서 있어서 금융회사의 개인정보에 대한 기술적‧관리적 보호체계의 실질적인 강화를 통한 금융사고 예방이 보다 중요하다고 할 수 있음 □ 한편 우리나라는 해외와 비교해보면 빅데이터 활용이 초기단계에 해당하지만, 인터넷전문은행 도입과 관련하여 빅데이터 활용에 대한 금융회사의 관심이 증가하고 있음 ○ 현재 빅데이터를 활성화하기 위해 「신용정보의 이용 및 보호에 관한 법률」 개정을 통해 개인신용정보의 개념을 보다 명확하게 정의하여 금융회사가 활용할 수 있는 비식별정보를 구분하는 방안이 추진되고 있음 □ 하지만 개인신용정보 중에서 식별화정보와 비식별화 정보를 명확히 구분하는 것이 현실적으로 쉽지 않으므로 빅데이터의 활성화를 위해 개인정보의 균형있는 보호와 활용이 필요한 시점임 ○ 특히 개인신용정보에서 개인을 식별할 수 없는 비식별정보를 구분하는 경우에도 동 정보의 이용 목적을 제한하고, 비식별정보의 재식별 방지를 위한 안전성 확보 방안 등을 마련할 필요가 있음',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Hyperparameters
#### Non-Default Hyperparameters
- `per_device_train_batch_size`: 64
- `learning_rate`: 3e-05
- `num_train_epochs`: 1
- `max_steps`: 50
- `warmup_ratio`: 0.05
- `fp16`: True
- `batch_sampler`: no_duplicates
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: no
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 64
- `per_device_eval_batch_size`: 8
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 1
- `eval_accumulation_steps`: None
- `torch_empty_cache_steps`: None
- `learning_rate`: 3e-05
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1.0
- `num_train_epochs`: 1
- `max_steps`: 50
- `lr_scheduler_type`: linear
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.05
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: False
- `fp16`: True
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: None
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: True
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: False
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: False
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `dispatch_batches`: None
- `split_batches`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `eval_on_start`: False
- `eval_use_gather_object`: False
- `batch_sampler`: no_duplicates
- `multi_dataset_batch_sampler`: proportional
</details>
### Training Logs
| Epoch | Step | Training Loss |
|:-----:|:----:|:-------------:|
| 0.004 | 1 | 1.1692 |
| 0.008 | 2 | 1.0241 |
| 0.012 | 3 | 1.1237 |
| 0.016 | 4 | 0.9547 |
| 0.02 | 5 | 1.0183 |
| 0.024 | 6 | 0.9393 |
| 0.028 | 7 | 0.6666 |
| 0.032 | 8 | 0.7058 |
| 0.036 | 9 | 0.6336 |
| 0.04 | 10 | 0.5752 |
| 0.044 | 11 | 0.6901 |
| 0.048 | 12 | 0.6699 |
| 0.052 | 13 | 0.6184 |
| 0.056 | 14 | 0.5948 |
| 0.06 | 15 | 0.6546 |
| 0.064 | 16 | 0.5846 |
| 0.068 | 17 | 0.5892 |
| 0.072 | 18 | 0.5819 |
| 0.076 | 19 | 0.5602 |
| 0.08 | 20 | 0.5515 |
| 0.084 | 21 | 0.5359 |
| 0.088 | 22 | 0.5599 |
| 0.092 | 23 | 0.5104 |
| 0.096 | 24 | 0.4943 |
| 0.1 | 25 | 0.564 |
| 0.104 | 26 | 0.5545 |
| 0.108 | 27 | 0.4937 |
| 0.112 | 28 | 0.5283 |
| 0.116 | 29 | 0.512 |
| 0.12 | 30 | 0.552 |
| 0.124 | 31 | 0.5417 |
| 0.128 | 32 | 0.4607 |
| 0.132 | 33 | 0.4281 |
| 0.136 | 34 | 0.4764 |
| 0.14 | 35 | 0.5736 |
| 0.144 | 36 | 0.5312 |
| 0.148 | 37 | 0.4723 |
| 0.152 | 38 | 0.5169 |
| 0.156 | 39 | 0.4849 |
| 0.16 | 40 | 0.5347 |
| 0.164 | 41 | 0.48 |
| 0.168 | 42 | 0.4745 |
| 0.172 | 43 | 0.5061 |
| 0.176 | 44 | 0.5438 |
| 0.18 | 45 | 0.4942 |
| 0.184 | 46 | 0.5486 |
| 0.188 | 47 | 0.475 |
| 0.192 | 48 | 0.5054 |
| 0.196 | 49 | 0.3898 |
| 0.2 | 50 | 0.4726 |
### Framework Versions
- Python: 3.10.12
- Sentence Transformers: 3.2.1
- Transformers: 4.44.2
- PyTorch: 2.3.1+cu121
- Accelerate: 1.1.1
- Datasets: 2.21.0
- Tokenizers: 0.19.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### CachedMultipleNegativesRankingLoss
```bibtex
@misc{gao2021scaling,
title={Scaling Deep Contrastive Learning Batch Size under Memory Limited Setup},
author={Luyu Gao and Yunyi Zhang and Jiawei Han and Jamie Callan},
year={2021},
eprint={2101.06983},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
Xenova/opus-mt-es-it | Xenova | translation | [
"transformers.js",
"onnx",
"marian",
"text2text-generation",
"translation",
"base_model:Helsinki-NLP/opus-mt-es-it",
"base_model:quantized:Helsinki-NLP/opus-mt-es-it",
"region:us"
] | 1,693,955,542,000 | 2024-10-08T13:42:05 | 57 | 1 | ---
base_model: Helsinki-NLP/opus-mt-es-it
library_name: transformers.js
pipeline_tag: translation
---
https://huggingface.co/Helsinki-NLP/opus-mt-es-it with ONNX weights to be compatible with Transformers.js.
Note: Having a separate repo for ONNX weights is intended to be a temporary solution until WebML gains more traction. If you would like to make your models web-ready, we recommend converting to ONNX using [🤗 Optimum](https://huggingface.co/docs/optimum/index) and structuring your repo like this one (with ONNX weights located in a subfolder named `onnx`). | [
"TRANSLATION"
] | Non_BioNLP |
TehranNLP-org/bert-large-mnli | TehranNLP-org | text-classification | [
"transformers",
"pytorch",
"bert",
"text-classification",
"generated_from_trainer",
"en",
"dataset:mnli",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,674,165,998,000 | 2023-01-19T22:22:28 | 116 | 0 | ---
datasets:
- mnli
language:
- en
license: apache-2.0
metrics:
- accuracy
tags:
- generated_from_trainer
model-index:
- name: '42'
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: MNLI
type: glue
args: mnli
metrics:
- type: accuracy
value: 0.8633723892002038
name: Accuracy
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# 42
This model is a fine-tuned version of [bert-large-uncased](https://huggingface.co/bert-large-uncased) on the MNLI dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8447
- Accuracy: 0.8634
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- distributed_type: not_parallel
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:--------:|
| 0.4274 | 1.0 | 12272 | 0.3892 | 0.8524 |
| 0.2844 | 2.0 | 24544 | 0.4079 | 0.8565 |
| 0.1589 | 3.0 | 36816 | 0.5033 | 0.8527 |
| 0.0877 | 4.0 | 49088 | 0.6624 | 0.8576 |
| 0.0426 | 5.0 | 61360 | 0.8447 | 0.8634 |
### Framework versions
- Transformers 4.17.0
- Pytorch 1.10.0+cu113
- Datasets 2.7.1
- Tokenizers 0.11.6
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
HARDYCHEN/text_summarization_finetuned | HARDYCHEN | text2text-generation | [
"transformers",
"tensorboard",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"base_model:Falconsai/text_summarization",
"base_model:finetune:Falconsai/text_summarization",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,714,016,412,000 | 2024-04-25T03:40:37 | 5 | 0 | ---
base_model: Falconsai/text_summarization
license: apache-2.0
metrics:
- rouge
tags:
- generated_from_trainer
model-index:
- name: text_summarization_finetuned
results: []
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# text_summarization_finetuned
This model is a fine-tuned version of [Falconsai/text_summarization](https://huggingface.co/Falconsai/text_summarization) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2709
- Rouge1: 0.0876
- Rouge2: 0.0826
- Rougel: 0.0876
- Rougelsum: 0.0876
- Gen Len: 19.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 4
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:------:|:------:|:------:|:---------:|:-------:|
| 0.3375 | 1.0 | 4000 | 0.2961 | 0.0876 | 0.0826 | 0.0876 | 0.0876 | 19.0 |
| 0.3046 | 2.0 | 8000 | 0.2776 | 0.0876 | 0.0826 | 0.0876 | 0.0876 | 19.0 |
| 0.2929 | 3.0 | 12000 | 0.2726 | 0.0876 | 0.0826 | 0.0876 | 0.0876 | 19.0 |
| 0.2915 | 4.0 | 16000 | 0.2709 | 0.0876 | 0.0826 | 0.0876 | 0.0876 | 19.0 |
### Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0
| [
"SUMMARIZATION"
] | Non_BioNLP |
SEBIS/legal_t5_small_multitask_cs_de | SEBIS | text2text-generation | [
"transformers",
"pytorch",
"jax",
"t5",
"text2text-generation",
"translation Cszech Deustch model",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,646,263,744,000 | 2021-06-23T10:50:44 | 175 | 0 | ---
datasets:
- dcep europarl jrc-acquis
language: Cszech Deustch
tags:
- translation Cszech Deustch model
widget:
- text: Postavení žen v ozbrojených konfliktech a jejich úloha při obnově zemí po
ukončení konfliktu a v demokratickém procesu v těchto zemích
---
# legal_t5_small_multitask_cs_de model
Model on translating legal text from Cszech to Deustch. It was first released in
[this repository](https://github.com/agemagician/LegalTrans). The model is parallely trained on the three parallel corpus with 42 language pair
from jrc-acquis, europarl and dcep along with the unsupervised task where the model followed the task of prediction in a masked language model.
## Model description
No pretraining is involved in case of legal_t5_small_multitask_cs_de model, rather the unsupervised task is added with all the translation task
to realize the multitask learning scenario.
## Intended uses & limitations
The model could be used for translation of legal texts from Cszech to Deustch.
### How to use
Here is how to use this model to translate legal text from Cszech to Deustch in PyTorch:
```python
from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline
pipeline = TranslationPipeline(
model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_multitask_cs_de"),
tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_multitask_cs_de", do_lower_case=False,
skip_special_tokens=True),
device=0
)
cs_text = "Postavení žen v ozbrojených konfliktech a jejich úloha při obnově zemí po ukončení konfliktu a v demokratickém procesu v těchto zemích"
pipeline([cs_text], max_length=512)
```
## Training data
The legal_t5_small_multitask_cs_de model (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 5 Million parallel texts.
## Training procedure
The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule.
### Preprocessing
An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model.
### Pretraining
## Evaluation results
When the model is used for translation test dataset, achieves the following results:
Test results :
| Model | BLEU score |
|:-----:|:-----:|
| legal_t5_small_multitask_cs_de | 43.145|
### BibTeX entry and citation info
> Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) | [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/)
| [
"TRANSLATION"
] | Non_BioNLP |
RichardErkhov/bertin-project_-_bertin-gpt-j-6B-alpaca-4bits | RichardErkhov | null | [
"safetensors",
"gptj",
"4-bit",
"bitsandbytes",
"region:us"
] | 1,731,426,189,000 | 2024-11-12T15:45:23 | 5 | 0 | ---
{}
---
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
bertin-gpt-j-6B-alpaca - bnb 4bits
- Model creator: https://huggingface.co/bertin-project/
- Original model: https://huggingface.co/bertin-project/bertin-gpt-j-6B-alpaca/
Original model description:
---
license: openrail
datasets:
- bertin-project/alpaca-spanish
library_name: transformers
language:
- es
pipeline_tag: text-generation
tags:
- alpaca
- ggml
widget:
- text: >-
A continuación hay una instrucción que describe una tarea. Escribe una
respuesta que complete adecuadamente lo que se pide.
### Instrucción: Escribe un correo electrónico dando la bienvenida a un
nuevo empleado llamado Manolo.
### Respuesta:
example_title: E-mail
- text: >-
A continuación hay una instrucción que describe una tarea. Escribe una
respuesta que complete adecuadamente lo que se pide.
### Instrucción: Cuéntame algo sobre las alpacas.
### Respuesta:
example_title: Alpacas
- text: >-
A continuación hay una instrucción que describe una tarea. Escribe una
respuesta que complete adecuadamente lo que se pide.
### Instrucción: Inventa una excusa creativa para decir que no tengo que ir
a la fiesta.
### Respuesta:
example_title: Excusa
---
# BERTIN-GPT-J-6B Alpaca
This is a [BERTIN GPT-J-6B](https://huggingface.co/bertin-project/bertin-gpt-j-6B) Spanish model fine-tuned on the [Spanish Alpaca](https://huggingface.co/datasets/bertin-project/alpaca-spanish) dataset.
## Usage
```python
from transformers import AutoModelForCausalLM, AutoTokenizer, GenerationConfig, pipeline
base_model = "bertin-project/bertin-gpt-j-6B-alpaca"
tokenizer = AutoTokenizer.from_pretrained(base_model)
model = AutoModelForCausalLM.from_pretrained(base_model).cuda()
```
For generation, we can either use `pipeline()` or the model's `.generate()` method. Remember that the prompt needs a **Spanish** template:
```python
# Generate responses
def generate(instruction, input=None):
if input:
prompt = f"""A continuación hay una instrucción que describe una tarea, junto con una entrada que proporciona más contexto. Escribe una respuesta que complete adecuadamente lo que se pide.
### Instrucción:
{instruction}
### Entrada:
{input}
### Respuesta:"""
else:
prompt = f""""A continuación hay una instrucción que describe una tarea. Escribe una respuesta que complete adecuadamente lo que se pide.
### Instrucción:
{instruction}
### Respuesta:
"""
inputs = tokenizer(prompt, return_tensors="pt")
input_ids = inputs["input_ids"].cuda()
generation_output = model.generate(
input_ids=input_ids,
generation_config=GenerationConfig(temperature=0.2, top_p=0.75, num_beams=4),
return_dict_in_generate=True,
output_scores=True,
max_new_tokens=256
)
for seq in generation_output.sequences:
output = tokenizer.decode(seq, skip_special_tokens=True)
print(output.split("### Respuesta:")[-1].strip())
generate("Escribe un correo electrónico dando la bienvenida a un nuevo empleado llamado Manolo.")
# Estimado Manolo,
#
# ¡Bienvenido a tu nuevo trabajo como Representante de Servicio al Cliente en nuestra empresa! Estamos emocionados de tenerte a bordo y esperamos que tengas un gran año trabajando con nosotros.
#
# En nombre de todos en esta empresa, queremos darte la bienvenida al equipo y desearte lo mejor en tus nuevas funciones.
#
# ¡Estamos ansiosos por escuchar tus historias y ayudarte a tener éxito en tu nuevo rol!
#
# Sinceramente,
# El equipo de Servicio al Cliente
```
## Data
The dataset is a translation to Spanish of [alpaca_data_cleaned.json](https://github.com/tloen/alpaca-lora/blob/main/alpaca_data_cleaned.json) (a clean version of the [Alpaca dataset made at Stanford](https://huggingface.co/datasets/tatsu-lab/alpaca)) using OpenAI's `gpt-3.5-turbo` model. We translated using a full-sample prompt instead of per strings, which resulted in more coherent tuples of `(instruction, input, output)` and costed around $60.0.
**This dataset cannot be used to create models that compete in any way with OpenAI.**
## Finetuning
To fine-tune the BERTIN GPT-J-6B model we used the code available on [BERTIN's fork of `mesh-transformer-jax`](https://github.com/bertin-project/mesh-transformer-jax/blob/master/prepare_dataset_alpaca.py), which provides code adapt an Alpaca dataset to finetune any GPT-J-6B model. We run finetuning for 3 epochs using sequence length of 2048 on a single TPUv3-8 for 3 hours on top of BERTIN GPT-J-6B.
## Example outputs
Stanford Alpaca:
```
Instruction: Inventa una excusa creativa para decir que no tengo que ir a la fiesta.
Response: I'm sorry, I can't make it to the party. I have to stay home and take care of my sick pet.
```
BERTIN Alpaca LoRA:
```
Instruction: Inventa una excusa creativa para decir que no tengo que ir a la fiesta.
Response: No puedo ir a la fiesta porque tengo que terminar un proyecto urgente.
```
BERTIN GPT-J-6B Alpaca:
```
Instruction: Inventa una excusa creativa para decir que no tengo que ir a la fiesta.
Response: Lo siento, pero tengo una cita inesperada a la que necesito asistir.
```
Stanford Alpaca:
```
Instruction: ¿Cómo puedo comenzar una carrera como científico de datos? Responde con una lista de pasos a seguir.
Response: To begin a career as a data scientist, there are a few steps to take. First, you should have a bachelor's degree in a STEM-related field, such as computer science, mathematics, statistics, or engineering. You should also have experience with programming languages such as Python, R, or Java. Additionally, it is helpful to have knowledge of machine learning, data visualization, and data analytics. Finally, it is important to stay up to date with the latest trends and technologies in the field.
```
BERTIN Alpaca LoRA:
```
Instruction: ¿Cómo puedo comenzar una carrera como científico de datos? Responde con una lista de pasos a seguir.
Response: Para comenzar una carrera como científico de datos, hay varios pasos a seguir. Primero, obtener una licenciatura en matemáticas, informática o ciencias de la computación. Luego, obtener una maestría o doctorado en ciencias de la computación, informática o alguna otra área relacionada. Finalmente, obtener experiencia en el campo trabajando en proyectos de investigación o desarrollando aplicaciones.
```
BERTIN GPT-J-6B Alpaca:
```
Instruction: ¿Cómo puedo comenzar una carrera como científico de datos? Responde con una lista de pasos a seguir.
Response: Comenzar una carrera como científico de datos requiere una sólida formación en matemáticas, estadística e informática. También es importante tener experiencia en análisis de datos, aprendizaje automático y visualización de datos. Para prepararse para una carrera como científico de datos, considere tomar cursos en estadística, aprendizaje automático, visualización de datos y otros temas relevantes. Además, asegúrese de obtener experiencia práctica trabajando en proyectos de análisis de datos o tomando roles de aprendiz de científico de datos.
```
You can test it using the eval notebook [here](https://colab.research.google.com/github/22-hours/cabrita/blob/main/notebooks/cabrita-lora.ipynb).
## References
- [LLaMA](https://ai.facebook.com/blog/large-language-model-llama-meta-ai/)
- [Stanford Alpaca](https://github.com/tatsu-lab/stanford_alpaca)
- [BERTIN Alpaca](https://huggingface.co/datasets/bertin-project/alpaca-spanish)
- [ChatGPT](https://openai.com/blog/chatgpt)
- [Hugging Face](https://huggingface.co/)
## Hardware Requirements
For training we have used a Google Cloud TPUv3-8 VM. For eval, you can use a T4.
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_bertin-project__bertin-gpt-j-6B-alpaca)
| Metric | Value |
|-----------------------|---------------------------|
| Avg. | 32.11 |
| ARC (25-shot) | 36.01 |
| HellaSwag (10-shot) | 54.3 |
| MMLU (5-shot) | 27.66 |
| TruthfulQA (0-shot) | 43.38 |
| Winogrande (5-shot) | 55.8 |
| GSM8K (5-shot) | 0.0 |
| DROP (3-shot) | 7.59 |
| [
"TRANSLATION"
] | Non_BioNLP |
RichardErkhov/rawsh_-_simpo-math-model-gguf | RichardErkhov | null | [
"gguf",
"arxiv:2401.08417",
"endpoints_compatible",
"region:us"
] | 1,741,879,944,000 | 2025-03-13T15:39:11 | 352 | 0 | ---
{}
---
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
simpo-math-model - GGUF
- Model creator: https://huggingface.co/rawsh/
- Original model: https://huggingface.co/rawsh/simpo-math-model/
| Name | Quant method | Size |
| ---- | ---- | ---- |
| [simpo-math-model.Q2_K.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q2_K.gguf) | Q2_K | 0.32GB |
| [simpo-math-model.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.IQ3_XS.gguf) | IQ3_XS | 0.32GB |
| [simpo-math-model.IQ3_S.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.IQ3_S.gguf) | IQ3_S | 0.32GB |
| [simpo-math-model.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q3_K_S.gguf) | Q3_K_S | 0.32GB |
| [simpo-math-model.IQ3_M.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.IQ3_M.gguf) | IQ3_M | 0.32GB |
| [simpo-math-model.Q3_K.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q3_K.gguf) | Q3_K | 0.33GB |
| [simpo-math-model.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q3_K_M.gguf) | Q3_K_M | 0.33GB |
| [simpo-math-model.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q3_K_L.gguf) | Q3_K_L | 0.34GB |
| [simpo-math-model.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.IQ4_XS.gguf) | IQ4_XS | 0.33GB |
| [simpo-math-model.Q4_0.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q4_0.gguf) | Q4_0 | 0.33GB |
| [simpo-math-model.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.IQ4_NL.gguf) | IQ4_NL | 0.33GB |
| [simpo-math-model.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q4_K_S.gguf) | Q4_K_S | 0.36GB |
| [simpo-math-model.Q4_K.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q4_K.gguf) | Q4_K | 0.37GB |
| [simpo-math-model.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q4_K_M.gguf) | Q4_K_M | 0.37GB |
| [simpo-math-model.Q4_1.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q4_1.gguf) | Q4_1 | 0.35GB |
| [simpo-math-model.Q5_0.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q5_0.gguf) | Q5_0 | 0.37GB |
| [simpo-math-model.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q5_K_S.gguf) | Q5_K_S | 0.38GB |
| [simpo-math-model.Q5_K.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q5_K.gguf) | Q5_K | 0.39GB |
| [simpo-math-model.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q5_K_M.gguf) | Q5_K_M | 0.39GB |
| [simpo-math-model.Q5_1.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q5_1.gguf) | Q5_1 | 0.39GB |
| [simpo-math-model.Q6_K.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q6_K.gguf) | Q6_K | 0.47GB |
| [simpo-math-model.Q8_0.gguf](https://huggingface.co/RichardErkhov/rawsh_-_simpo-math-model-gguf/blob/main/simpo-math-model.Q8_0.gguf) | Q8_0 | 0.49GB |
Original model description:
---
base_model: rawsh/mirrorqwen2.5-0.5b-SFT
library_name: transformers
model_name: simpo-math-model
tags:
- generated_from_trainer
- trl
- cpo
- unsloth
licence: license
---
# Model Card for simpo-math-model
This model is a fine-tuned version of [rawsh/mirrorqwen2.5-0.5b-SFT](https://huggingface.co/rawsh/mirrorqwen2.5-0.5b-SFT).
It has been trained using [TRL](https://github.com/huggingface/trl).
## Quick start
```python
from transformers import pipeline
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="rawsh/simpo-math-model", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```
## Training procedure
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/dankgpt/simpo-training/runs/q29stpw6)
This model was trained with CPO, a method introduced in [Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation](https://huggingface.co/papers/2401.08417).
### Framework versions
- TRL: 0.12.0
- Transformers: 4.46.2
- Pytorch: 2.4.1
- Datasets: 3.1.0
- Tokenizers: 0.20.3
## Citations
Cite CPO as:
```bibtex
@inproceedings{xu2024contrastive,
title = {{Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation}},
author = {Haoran Xu and Amr Sharaf and Yunmo Chen and Weiting Tan and Lingfeng Shen and Benjamin Van Durme and Kenton Murray and Young Jin Kim},
year = 2024,
booktitle = {Forty-first International Conference on Machine Learning, {ICML} 2024, Vienna, Austria, July 21-27, 2024},
publisher = {OpenReview.net},
url = {https://openreview.net/forum?id=51iwkioZpn}
}
```
Cite TRL as:
```bibtex
@misc{vonwerra2022trl,
title = {{TRL: Transformer Reinforcement Learning}},
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
year = 2020,
journal = {GitHub repository},
publisher = {GitHub},
howpublished = {\url{https://github.com/huggingface/trl}}
}
```
| [
"TRANSLATION"
] | Non_BioNLP |
mrapacz/interlinear-en-philta-emb-auto-diacritics-bh | mrapacz | text2text-generation | [
"transformers",
"pytorch",
"morph-t5-auto",
"text2text-generation",
"en",
"dataset:mrapacz/greek-interlinear-translations",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,739,017,600,000 | 2025-02-21T21:32:33 | 45 | 0 | ---
base_model:
- PhilTa
datasets:
- mrapacz/greek-interlinear-translations
language:
- en
library_name: transformers
license: cc-by-sa-4.0
metrics:
- bleu
---
# Model Card for Ancient Greek to English Interlinear Translation Model
This model performs interlinear translation from Ancient Greek to English, maintaining word-level alignment between source and target texts.
You can find the source code used for training this and other models trained as part of this project in the [GitHub repository](https://github.com/mrapacz/loreslm-interlinear-translation).
## Model Details
### Model Description
- **Developed By:** Maciej Rapacz, AGH University of Kraków
- **Model Type:** MorphT5AutoForConditionalGeneration
- **Base Model:** PhilTa
- **Tokenizer:** PhilTa
- **Language(s):** Ancient Greek (source) → English (target)
- **License:** CC BY-NC-SA 4.0
- **Tag Set:** BH (Bible Hub)
- **Text Preprocessing:** Diacritics
- **Morphological Encoding:** emb-auto
### Model Performance
- **BLEU Score:** 60.40
- **SemScore:** 0.89
### Model Sources
- **Repository:** https://github.com/mrapacz/loreslm-interlinear-translation
- **Paper:** https://aclanthology.org/2025.loreslm-1.11/
## Usage Example
> **Note**: This model uses a modification of T5-family models that includes dedicated embedding layers for encoding morphological information. To load these models, install the [morpht5](https://github.com/mrapacz/loreslm-interlinear-translation/blob/master/morpht5/README.md) package:
> ```bash
> pip install morpht5
> ```
```python
>>> from morpht5 import MorphT5AutoForConditionalGeneration, MorphT5Tokenizer
>>> text = ['Λέγει', 'αὐτῷ', 'ὁ', 'Ἰησοῦς', 'Ἔγειρε', 'ἆρον', 'τὸν', 'κράβαττόν', 'σου', 'καὶ', 'περιπάτει']
>>> tags = ['V-PIA-3S', 'PPro-DM3S', 'Art-NMS', 'N-NMS', 'V-PMA-2S', 'V-AMA-2S', 'Art-AMS', 'N-AMS', 'PPro-G2S', 'Conj', 'V-PMA-2S']
>>> tokenizer = MorphT5Tokenizer.from_pretrained("mrapacz/interlinear-en-philta-emb-auto-diacritics-bh")
>>> inputs = tokenizer(
text=text,
morph_tags=tags,
return_tensors="pt"
)
>>> model = MorphT5AutoForConditionalGeneration.from_pretrained("mrapacz/interlinear-en-philta-emb-auto-diacritics-bh")
>>> outputs = model.generate(
**inputs,
max_new_tokens=100,
early_stopping=True,
)
>>> decoded = tokenizer.decode(outputs[0], skip_special_tokens=True, keep_block_separator=True)
>>> decoded = decoded.replace(tokenizer.target_block_separator_token, " | ")
>>> decoded
'says | to him | - | jesus | arise | take up | the | mat | of you | and | walk'
```
## Citation
If you use this model, please cite the following paper:
```
@inproceedings{rapacz-smywinski-pohl-2025-low,
title = "Low-Resource Interlinear Translation: Morphology-Enhanced Neural Models for {A}ncient {G}reek",
author = "Rapacz, Maciej and
Smywi{\'n}ski-Pohl, Aleksander",
editor = "Hettiarachchi, Hansi and
Ranasinghe, Tharindu and
Rayson, Paul and
Mitkov, Ruslan and
Gaber, Mohamed and
Premasiri, Damith and
Tan, Fiona Anting and
Uyangodage, Lasitha",
booktitle = "Proceedings of the First Workshop on Language Models for Low-Resource Languages",
month = jan,
year = "2025",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.loreslm-1.11/",
pages = "145--165",
abstract = "Contemporary machine translation systems prioritize fluent, natural-sounding output with flexible word ordering. In contrast, interlinear translation maintains the source text`s syntactic structure by aligning target language words directly beneath their source counterparts. Despite its importance in classical scholarship, automated approaches to interlinear translation remain understudied. We evaluated neural interlinear translation from Ancient Greek to English and Polish using four transformer-based models: two Ancient Greek-specialized (GreTa and PhilTa) and two general-purpose multilingual models (mT5-base and mT5-large). Our approach introduces novel morphological embedding layers and evaluates text preprocessing and tag set selection across 144 experimental configurations using a word-aligned parallel corpus of the Greek New Testament. Results show that morphological features through dedicated embedding layers significantly enhance translation quality, improving BLEU scores by 35{\%} (44.67 {\textrightarrow} 60.40) for English and 38{\%} (42.92 {\textrightarrow} 59.33) for Polish compared to baseline models. PhilTa achieves state-of-the-art performance for English, while mT5-large does so for Polish. Notably, PhilTa maintains stable performance using only 10{\%} of training data. Our findings challenge the assumption that modern neural architectures cannot benefit from explicit morphological annotations. While preprocessing strategies and tag set selection show minimal impact, the substantial gains from morphological embeddings demonstrate their value in low-resource scenarios."
}
``` | [
"TRANSLATION"
] | Non_BioNLP |
ymoslem/whisper-medium-ga2en-v6.3.1-8k-r | ymoslem | automatic-speech-recognition | [
"transformers",
"tensorboard",
"safetensors",
"whisper",
"automatic-speech-recognition",
"generated_from_trainer",
"ga",
"en",
"dataset:ymoslem/IWSLT2023-GA-EN",
"dataset:ymoslem/FLEURS-GA-EN",
"dataset:ymoslem/BitesizeIrish-GA-EN",
"dataset:ymoslem/SpokenWords-GA-EN-MTed",
"dataset:ymoslem/Tatoeba-Speech-Irish",
"dataset:ymoslem/Wikimedia-Speech-Irish",
"dataset:ymoslem/EUbookshop-Speech-Irish",
"base_model:openai/whisper-medium",
"base_model:finetune:openai/whisper-medium",
"license:apache-2.0",
"model-index",
"endpoints_compatible",
"region:us"
] | 1,719,016,140,000 | 2025-03-15T11:12:14 | 36 | 1 | ---
base_model: openai/whisper-medium
datasets:
- ymoslem/IWSLT2023-GA-EN
- ymoslem/FLEURS-GA-EN
- ymoslem/BitesizeIrish-GA-EN
- ymoslem/SpokenWords-GA-EN-MTed
- ymoslem/Tatoeba-Speech-Irish
- ymoslem/Wikimedia-Speech-Irish
- ymoslem/EUbookshop-Speech-Irish
language:
- ga
- en
license: apache-2.0
metrics:
- bleu
- wer
tags:
- generated_from_trainer
model-index:
- name: Whisper Medium GA-EN Speech Translation
results:
- task:
type: automatic-speech-recognition
name: Automatic Speech Recognition
dataset:
name: IWSLT-2023, FLEURS, BiteSize, SpokenWords, Tatoeba, Wikimedia, and EUbookshop
type: ymoslem/IWSLT2023-GA-EN
metrics:
- type: bleu
value: 32.0
name: Bleu
- type: wer
value: 66.77172444844665
name: Wer
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# Whisper Medium GA-EN Speech Translation
This model is a fine-tuned version of [openai/whisper-medium](https://huggingface.co/openai/whisper-medium) on the IWSLT-2023, FLEURS, BiteSize, SpokenWords, Tatoeba, Wikimedia, and EUbookshop dataset.
It achieves the following results on the evaluation set:
- Loss: 1.1067
- Bleu: 32.0
- Chrf: 52.48
- Wer: 66.7717
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.03
- training_steps: 8000
- mixed_precision_training: Native AMP
### Training results
| Training Loss | Epoch | Step | Bleu | Chrf | Validation Loss | Wer |
|:-------------:|:------:|:----:|:-----:|:-----:|:---------------:|:--------:|
| 2.5219 | 0.0138 | 100 | 0.44 | 10.48 | 2.1106 | 107.2490 |
| 2.4608 | 0.0276 | 200 | 3.3 | 20.43 | 2.1816 | 179.1535 |
| 2.3008 | 0.0414 | 300 | 3.66 | 21.59 | 2.0587 | 206.4836 |
| 2.2095 | 0.0552 | 400 | 8.79 | 27.66 | 1.9459 | 100.3602 |
| 2.0454 | 0.0690 | 500 | 8.14 | 27.36 | 1.8681 | 122.1522 |
| 1.9937 | 0.0828 | 600 | 11.05 | 30.26 | 1.8717 | 97.2535 |
| 1.868 | 0.0966 | 700 | 9.14 | 29.03 | 1.7917 | 129.0410 |
| 1.9924 | 0.1103 | 800 | 12.62 | 33.2 | 1.7170 | 89.6443 |
| 1.8646 | 0.1241 | 900 | 11.98 | 30.77 | 1.7252 | 97.8838 |
| 1.7644 | 0.1379 | 1000 | 10.87 | 31.0 | 1.6832 | 109.1851 |
| 1.692 | 0.1517 | 1100 | 13.05 | 34.46 | 1.6837 | 93.3814 |
| 1.7044 | 0.1655 | 1200 | 20.95 | 37.42 | 1.5527 | 75.2364 |
| 1.6824 | 0.1793 | 1300 | 14.91 | 35.56 | 1.5611 | 92.6159 |
| 1.6557 | 0.1931 | 1400 | 14.0 | 36.54 | 1.5554 | 99.8199 |
| 1.5456 | 0.2069 | 1500 | 19.72 | 39.81 | 1.5058 | 83.5660 |
| 1.3755 | 0.2207 | 1600 | 18.04 | 37.95 | 1.5039 | 82.9806 |
| 1.3959 | 0.2345 | 1700 | 17.01 | 39.5 | 1.4374 | 85.2319 |
| 1.5012 | 0.2483 | 1800 | 14.93 | 39.24 | 1.4242 | 114.4079 |
| 1.4278 | 0.2621 | 1900 | 23.85 | 42.69 | 1.3904 | 73.0302 |
| 1.3285 | 0.2759 | 2000 | 17.7 | 37.23 | 1.4493 | 83.8811 |
| 1.2655 | 0.2897 | 2100 | 20.1 | 40.32 | 1.3661 | 79.7839 |
| 1.2074 | 0.3034 | 2200 | 24.45 | 43.79 | 1.3387 | 72.9851 |
| 1.1893 | 0.3172 | 2300 | 21.45 | 42.61 | 1.3308 | 82.3953 |
| 1.1236 | 0.3310 | 2400 | 22.77 | 44.17 | 1.3050 | 77.3075 |
| 1.0934 | 0.3448 | 2500 | 25.54 | 46.32 | 1.2793 | 72.2647 |
| 1.06 | 0.3586 | 2600 | 28.27 | 47.32 | 1.2396 | 65.6911 |
| 1.0327 | 0.3724 | 2700 | 28.45 | 47.01 | 1.2577 | 67.3570 |
| 1.1623 | 0.3862 | 2800 | 24.54 | 47.43 | 1.2194 | 73.6155 |
| 1.0215 | 0.4 | 2900 | 27.4 | 49.6 | 1.2039 | 69.2481 |
| 0.9185 | 0.4138 | 3000 | 27.04 | 49.24 | 1.1724 | 67.8973 |
| 0.9003 | 0.4276 | 3100 | 31.08 | 50.11 | 1.1674 | 63.8001 |
| 0.9839 | 0.4414 | 3200 | 30.24 | 50.63 | 1.1580 | 64.5655 |
| 0.9396 | 0.4552 | 3300 | 30.79 | 51.72 | 1.1202 | 64.9257 |
| 0.9051 | 0.4690 | 3400 | 30.34 | 53.08 | 1.1180 | 66.4566 |
| 0.8621 | 0.4828 | 3500 | 33.3 | 53.86 | 1.1042 | 60.7834 |
| 0.8236 | 0.4966 | 3600 | 32.77 | 53.21 | 1.1070 | 62.0441 |
| 0.829 | 0.5103 | 3700 | 32.49 | 54.21 | 1.0771 | 62.5844 |
| 0.8375 | 0.5241 | 3800 | 32.27 | 53.98 | 1.0780 | 63.0797 |
| 0.8206 | 0.5379 | 3900 | 33.26 | 55.07 | 1.0615 | 61.6389 |
| 0.8059 | 0.5517 | 4000 | 33.24 | 55.16 | 1.0552 | 61.5038 |
| 0.9133 | 0.5655 | 4100 | 1.2218| 29.38 | 49.22 | 66.0964 |
| 1.051 | 0.5793 | 4200 | 1.2304| 25.12 | 46.01 | 71.8145 |
| 0.954 | 0.5931 | 4300 | 1.2501| 25.47 | 45.88 | 75.3715 |
| 0.939 | 0.6069 | 4400 | 1.2204| 29.19 | 47.63 | 66.9068 |
| 0.9887 | 0.6207 | 4500 | 1.2099| 27.99 | 47.01 | 67.7172 |
| 1.0044 | 0.6345 | 4600 | 1.2080| 23.77 | 45.33 | 73.3904 |
| 0.9881 | 0.6483 | 4700 | 1.2188| 26.46 | 47.36 | 68.5277 |
| 0.9674 | 0.6621 | 4800 | 1.2296| 26.11 | 45.92 | 68.3026 |
| 0.8845 | 0.6759 | 4900 | 1.2347| 27.3 | 46.08 | 68.0324 |
| 0.8297 | 0.6897 | 5000 | 1.2108| 29.48 | 48.96 | 64.6105 |
| 0.9065 | 0.7034 | 5100 | 1.1873| 29.81 | 49.94 | 64.2503 |
| 0.8096 | 0.7172 | 5200 | 1.2122| 28.5 | 46.93 | 66.2314 |
| 0.8077 | 0.7310 | 5300 | 1.1945| 29.26 | 48.21 | 64.4755 |
| 0.8227 | 0.7448 | 5400 | 1.2310| 26.82 | 48.43 | 71.4093 |
| 0.7587 | 0.7586 | 5500 | 1.2067| 29.45 | 49.03 | 65.3309 |
| 0.7206 | 0.7724 | 5600 | 1.2114| 29.89 | 49.33 | 65.5561 |
| 0.8088 | 0.7862 | 5700 | 1.1689| 31.88 | 51.4 | 64.2954 |
| 0.693 | 0.8 | 5800 | 1.1644| 27.23 | 48.11 | 68.7078 |
| 0.7099 | 0.8138 | 5900 | 1.1852| 31.01 | 49.42 | 63.3949 |
| 0.7564 | 0.8276 | 6000 | 1.1554| 28.3 | 50.34 | 71.0941 |
| 0.584 | 0.8414 | 6100 | 1.1566| 34.79 | 51.69 | 59.0725 |
| 0.6817 | 0.8552 | 6200 | 1.1245| 34.08 | 51.95 | 59.8829 |
| 0.5968 | 0.8690 | 6300 | 1.1475| 32.4 | 51.59 | 62.9896 |
| 0.6092 | 0.8828 | 6400 | 1.1250| 32.83 | 52.82 | 62.5844 |
| 0.6325 | 0.8966 | 6500 | 1.1108| 29.29 | 51.68 | 69.1130 |
| 0.6002 | 0.9103 | 6600 | 1.0993| 27.64 | 52.7 | 71.0941 |
| 0.6247 | 0.9241 | 6700 | 1.0898| 28.39 | 52.4 | 68.3026 |
| 0.6257 | 0.9379 | 6800 | 1.0863| 28.54 | 52.33 | 70.9140 |
| 0.6719 | 0.9517 | 6900 | 1.0891| 31.43 | 53.53 | 66.1414 |
| 0.4994 | 0.9655 | 7000 | 1.1066| 33.81 | 52.77 | 61.0986 |
| 0.5469 | 0.9793 | 7100 | 1.0891| 30.52 | 53.13 | 67.3570 |
| 0.6031 | 0.9931 | 7200 | 1.0933| 33.16 | 54.03 | 62.1792 |
| 0.2469 | 1.0069 | 7300 | 1.1426| 33.76 | 52.38 | 62.8546 |
| 0.2572 | 1.0207 | 7400 | 1.1292| 33.16 | 51.71 | 64.8807 |
| 0.2762 | 1.0345 | 7500 | 1.1090| 34.76 | 54.28 | 60.7384 |
| 0.2332 | 1.0483 | 7600 | 1.1073| 30.95 | 52.28 | 66.1864 |
| 0.2069 | 1.0621 | 7700 | 1.0999| 32.39 | 53.08 | 65.5561 |
| 0.2417 | 1.0759 | 7800 | 1.1008| 31.3 | 53.87 | 65.1058 |
| 0.2403 | 1.0897 | 7900 | 1.1053| 32.18 | 53.3 | 66.4566 |
| 0.208 | 1.1034 | 8000 | 1.1067| 32.0 | 52.48 | 66.7717 |
### Framework versions
- Transformers 4.41.2
- Pytorch 2.2.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
## Citation
```
@inproceedings{moslem-2024-leveraging,
title = "Leveraging Synthetic Audio Data for End-to-End Low-Resource Speech Translation",
author = "Moslem, Yasmin",
booktitle = "Proceedings of the 21st International Conference on Spoken Language Translation (IWSLT 2024)",
month = aug,
year = "2024",
address = "Bangkok, Thailand (in-person and online)",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.iwslt-1.31/",
doi = "10.18653/v1/2024.iwslt-1.31",
pages = "265--273",
abstract = "This paper describes our system submission to the International Conference on Spoken Language Translation (IWSLT 2024) for Irish-to-English speech translation. We built end-to-end systems based on Whisper, and employed a number of data augmentation techniques, such as speech back-translation and noise augmentation. We investigate the effect of using synthetic audio data and discuss several methods for enriching signal diversity."
}
``` | [
"TRANSLATION"
] | Non_BioNLP |
gaudi/opus-mt-fr-ny-ctranslate2 | gaudi | translation | [
"transformers",
"marian",
"ctranslate2",
"translation",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | 1,721,663,939,000 | 2024-10-19T04:38:59 | 7 | 0 | ---
license: apache-2.0
tags:
- ctranslate2
- translation
---
# Repository General Information
## Inspired by and derived from the work of [Helsinki-NLP](https://huggingface.co/Helsinki-NLP), [CTranslate2](https://github.com/OpenNMT/CTranslate2), and [michaelfeil](https://huggingface.co/michaelfeil)!
- Link to Original Model ([Helsinki-NLP](https://huggingface.co/Helsinki-NLP)): [Model Link](https://huggingface.co/Helsinki-NLP/opus-mt-fr-ny)
- This respository was based on the work of [CTranslate2](https://github.com/OpenNMT/CTranslate2).
- This repository was based on the work of [michaelfeil](https://huggingface.co/michaelfeil).
# What is CTranslate2?
[CTranslate2](https://opennmt.net/CTranslate2/) is a C++ and Python library for efficient inference with Transformer models.
CTranslate2 implements a custom runtime that applies many performance optimization techniques such as weights quantization, layers fusion, batch reordering, etc., to accelerate and reduce the memory usage of Transformer models on CPU and GPU.
CTranslate2 is one of the most performant ways of hosting translation models at scale. Current supported models include:
- Encoder-decoder models: Transformer base/big, M2M-100, NLLB, BART, mBART, Pegasus, T5, Whisper
- Decoder-only models: GPT-2, GPT-J, GPT-NeoX, OPT, BLOOM, MPT, Llama, Mistral, Gemma, CodeGen, GPTBigCode, Falcon
- Encoder-only models: BERT, DistilBERT, XLM-RoBERTa
The project is production-oriented and comes with backward compatibility guarantees, but it also includes experimental features related to model compression and inference acceleration.
# CTranslate2 Benchmarks
Please note that the results presented below are only valid for the configuration used during this benchmark: absolute and relative performance may change with different settings. Tested against `newstest2014` (En -> De) dataset.
The benchmark reports the number of target tokens generated per second (higher is better). The results are aggregated over multiple runs. See the benchmark scripts for more details and reproduce these numbers.
Please note that the results presented below are only valid for the configuration used during this benchmark: absolute and relative performance may change with different settings.
## CPU Benchmarks for Generic Opus-MT Models
| Library | Tokens per Second | Max Memory Usage | BLEU |
| :----: | :----: | :----: | :----: |
| Transformers 4.26.1 (with PyTorch 1.13.1) | 147.3 | 2332MB | 27.90 |
| Marian 1.11.0 (int16) | 330.2 | 5901MB | 27.65 |
| Marian 1.11.0 (int8) | 355.8 | 4763MB | 27.27 |
| CTranslate2 3.6.0 (int16) | 596.1 | 660MB | 27.53 |
| CTranslate2 3.6.0 (int8) | 696.1 | 516MB | 27.65 |
## GPU Benchmarks for Generic Opus-MT Models
| Library | Tokens per Second | Max GPU Memory Usage | Max Memory Usage | BLEU |
| :----: | :----: | :----: | :----: | :----: |
| Transformers 4.26.1 (with PyTorch 1.13.1) | 1022.9 | 4097MB | 2109MB | 27.90 |
| Marian 1.11.0 (float16) | 3962.4 | 3239MB | 1976MB | 27.94 |
| CTranslate2 3.6.0 (float16) | 9296.7 | 909MB | 814MB | 27.9 |
| CTranslate2 3.6.0 (int8 + float16) | 8362.7 | 813MB | 766MB | 27.9 |
`Executed with 4 threads on a c5.2xlarge Amazon EC2 instance equipped with an Intel(R) Xeon(R) Platinum 8275CL CPU.`
**Source to benchmark information can be found [here](https://github.com/OpenNMT/CTranslate2).**<br />
**Original model BLEU scores can be found [here](https://huggingface.co/Helsinki-NLP/opus-mt-fr-ny).**
## Internal Benchmarks
Internal testing on our end showed **inference times reduced by 6x-10x** on average compared the vanilla checkpoints using the *transformers* library. A **slight reduction on BLEU scores (~5%)** was also identified in comparison to the vanilla checkpoints with a few exceptions. This is likely due to several factors, one being the quantization applied. Further testing is needed from our end to better assess the reduction in translation quality. The command used to compile the vanilla checkpoint into a CTranslate2 model can be found below. Modifying this command can yield differing balances between inferencing performance and translation quality.
# CTranslate2 Installation
```bash
pip install hf-hub-ctranslate2>=1.0.0 ctranslate2>=3.13.0
```
### ct2-transformers-converter Command Used:
```bash
ct2-transformers-converter --model Helsinki-NLP/opus-mt-fr-ny --output_dir ./ctranslate2/opus-mt-fr-ny-ctranslate2 --force --copy_files README.md generation_config.json tokenizer_config.json vocab.json source.spm .gitattributes target.spm --quantization float16
```
# CTranslate2 Converted Checkpoint Information:
**Compatible With:**
- [ctranslate2](https://github.com/OpenNMT/CTranslate2)
- [hf-hub-ctranslate2](https://github.com/michaelfeil/hf-hub-ctranslate2)
**Compute Type:**
- `compute_type=int8_float16` for `device="cuda"`
- `compute_type=int8` for `device="cpu"`
# Sample Code - ctranslate2
#### Clone the repository to the working directory or wherever you wish to store the model artifacts. ####
```bash
git clone https://huggingface.co/gaudi/opus-mt-fr-ny-ctranslate2
```
#### Take the python code below and update the 'model_dir' variable to the location of the cloned repository. ####
```python
from ctranslate2 import Translator
import transformers
model_dir = "./opus-mt-fr-ny-ctranslate2" # Path to model directory.
translator = Translator(
model_path=model_dir,
device="cuda", # cpu, cuda, or auto.
inter_threads=1, # Maximum number of parallel translations.
intra_threads=4, # Number of OpenMP threads per translator.
compute_type="int8_float16", # int8 for cpu or int8_float16 for cuda.
)
tokenizer = transformers.AutoTokenizer.from_pretrained(model_dir)
source = tokenizer.convert_ids_to_tokens(tokenizer.encode("XXXXXX, XXX XX XXXXXX."))
results = translator.translate_batch([source])
target = results[0].hypotheses[0]
print(tokenizer.decode(tokenizer.convert_tokens_to_ids(target)))
```
# Sample Code - hf-hub-ctranslate2
**Derived From [michaelfeil](https://huggingface.co/michaelfeil):**
```python
from hf_hub_ctranslate2 import TranslatorCT2fromHfHub, GeneratorCT2fromHfHub
from transformers import AutoTokenizer
model_name = "gaudi/opus-mt-fr-ny-ctranslate2"
model = TranslatorCT2fromHfHub(
model_name_or_path=model_name,
device="cuda",
compute_type="int8_float16",
tokenizer=AutoTokenizer.from_pretrained(model_name)
)
outputs = model.generate(
text=["XXX XX XXX XXXXXXX XXXX?", "XX XX XXXX XX XXX!"],
)
print(outputs)
```
# License and other remarks:
License conditions are intended to be idential to [original huggingface repository](https://huggingface.co/Helsinki-NLP/opus-mt-fr-ny) by Helsinki-NLP.
| [
"TRANSLATION"
] | Non_BioNLP |
RayNguyent/finetuning-sentiment-model | RayNguyent | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:imdb",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,690,603,011,000 | 2023-07-29T10:46:52 | 13 | 0 | ---
base_model: distilbert-base-uncased
datasets:
- imdb
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: finetuning-sentiment-model
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: imdb
type: imdb
config: plain_text
split: test
args: plain_text
metrics:
- type: accuracy
value: 1.0
name: Accuracy
- type: f1
value: 0.0
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# finetuning-sentiment-model
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0819
- Accuracy: 1.0
- F1: 0.0
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
### Framework versions
- Transformers 4.31.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.1
- Tokenizers 0.13.3
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
cmgx/Snowflake-ATM-Avg-v2 | cmgx | sentence-similarity | [
"sentence-transformers",
"safetensors",
"bert",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:800",
"loss:MatryoshkaLoss",
"loss:CustomContrastiveLoss",
"en",
"arxiv:1908.10084",
"arxiv:2205.13147",
"base_model:Snowflake/snowflake-arctic-embed-m-v1.5",
"base_model:finetune:Snowflake/snowflake-arctic-embed-m-v1.5",
"model-index",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,727,995,994,000 | 2024-10-03T23:01:03 | 0 | 0 | ---
base_model: Snowflake/snowflake-arctic-embed-m-v1.5
datasets: []
language:
- en
library_name: sentence-transformers
metrics:
- cosine_accuracy
- dot_accuracy
- manhattan_accuracy
- euclidean_accuracy
- max_accuracy
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:800
- loss:MatryoshkaLoss
- loss:CustomContrastiveLoss
widget:
- source_sentence: Hi
sentences:
- 'UNITED STATES SECURITIES AND EXCHANGE COMMISSION Washington, D.C. 20549 FORM
8-K CURRENT REPORT Pursuant to Section 13 or 15(d) of the Securities Exchange
Act of 1934 Date of Report (Date of earliest event reported): August 21, 2024
( August 16, 2024 ) SinglePoint Inc.
(Exact name of registrant as specified in
its charter) Nevada 000-53425 26-1240905 (State or other jurisdiction of
Incorporation) (Commission File Number) (IRS Employer Identification No.)
3104
E Camelback Rd #2137 Phoenix , AZ 85016 (Address of principal executive
offices) (Zip Code) Registrant’s telephone number, including area code: ( 888
) 682-7464 Not Applicable (Former name or former address, if changed since
last report.)Check the appropriate box below if the Form 8 K filing is
intended to simultaneously satisfy the filing obligation of the registrant
under any of the following provisions ( see General Instruction A.2.
below): ☐
Written communications pursuant to Rule 425 under the Securities Act (17 CFR
230.425)☐ Soliciting material pursuant to Rule 14a-12 under the Exchange Act
(17 CFR 240.14a-12)☐ Pre commencement communications pursuant to Rule
14d-2(b) under the Exchange Act (17 CFR 240.14d-2(b))☐ Pre commencement
communications pursuant to Rule 13e-4(c) under the Exchange Act (17 CFR
240.13e-4(c))Securities registered pursuant to Section 12(b) of the Act:
Title of each class Trading Symbol(s) Name of each exchange on which
registered common stock, par value $0.0001 per share SING Cboe BZX Exchange,
Inc. Indicate by check mark whether the registrant is an emerging growth
company as defined in Rule 405 of the Securities Act of 1933 (§230.405 of this
chapter) or Rule 12b-2 of the Securities Exchange Act of 1934 (§240.12b-2 of
this chapter).Emerging growth company ☐ If an emerging growth company,
indicate by check mark if the registrant has elected not to use the extended
transition period for complying with any new or revised financial accounting
standards provided pursuant to Section 13(a) of the Exchange Act.'
- 'Between April 1, 2024, and June 30, 2024, 39,167 restricted stock
units of the Company issued under the 2022 Plan were canceled by the Board of
Directors.On April 2, 2024, the Company issued 5,000 shares of restricted
common stock to its officer under the 2022 Plan.On April 23, 2024, the
Company issued 159,167 shares of restricted common stock to its officers and
directors under the 2022 Plan in exchange for cancellation of all stock
options and restricted stock units held by officers and directors of the
Company.On April 30, 2024, 3,750 restricted stock units held by consultant
were converted into 3,750 shares of common stock of the Company in connection
with the services provided by the consultant.On June 4, 2024, MGO issued a
total of 182,868 shares of the Company’s restricted common stock to directors
and officers of the Company pursuant to the 2022 Plan.The stock options,
restricted stock units, and the common stock issued or issuable upon the
exercise of such options and restricted stock units as described in this
section were issued pursuant to written compensatory plans or arrangements
with our employees, consultants, officers and directors, in reliance on the
exemption from the registration requirements of the Securities Act provided by
Rule 701 promulgated under the Securities Act or the exemption set forth in
Section 4(a)(2) under the Securities Act and Regulation D promulgated
thereunder relative to transactions by an issuer not involving any public
offering.
All recipients either received adequate information about us or had
access, through employment or other relationships, to such information.ITEM
3.'
- 'On August 17, 2022, a
registration statement (the “First Registration Statement”) was declared
effective to cover the resale of up to 633,333 shares of the Company’s common
stock comprised of (i) the 32,846 initial commitment shares, and (ii) up to
600,486 that the Company has reserved for issuance and sale to Lincoln Park
under the 2022 Purchase Agreement from time to time from and after the date of
the prospectus.
The Company sold approximately 527,166 shares under the First
Registration Statement.On August 18, 2023, a second registration statement
(the “Second Registration Statement”) was declared effective to cover the
resale of up to an additional 1,500,000 shares of the Company’s common stock
that the Company reserved for issuance and sale to Lincoln Park under the 2022
Purchase Agreement from time to time.
The Company sold 150,000 shares under
the Second Registration Statement.
The Company cannot sell more shares than
registered under the Second Registration Statement under the 2022 Purchase
Agreement without registering additional shares.'
- source_sentence: Hi
sentences:
- 'Although the
Company has filed the Prospectus Supplement with the Securities and Exchange
Commission, the Company has no obligation to sell any Shares under the Equity
Distribution Agreements, and may at any time suspend the offering of Shares
under the Equity Distribution Agreements.
Actual sales will depend on a
variety of factors to be determined by the Company from time to time,
including, among others, market conditions, the trading price of the Shares
and determinations by the Company of its need for, and the appropriate sources
of, additional capital.The Equity Distribution Agreements contain customary
representations, warranties and agreements of the Company, conditions to
closing, indemnification rights and obligations of the parties and termination
provisions.The foregoing description is only a summary of the material
provisions of the Equity Distribution Agreements and does not purport to be
complete and is qualified in its entirety by reference to the full text of the
Form of Equity Distribution Agreements, filed as Exhibit 10.1 to this Current
Report on Form 8-K and incorporated by reference herein.A copy of the opinion
of Miles & Stockbridge P.C. relating to the legality of the issuance and sale
of the Shares pursuant to the Prospectus is attached as Exhibit 5.1 hereto.1
This Current Report on Form 8-K shall not constitute an offer to sell or a
solicitation of an offer to buy any securities, nor shall there be any sale of
these securities in any state or jurisdiction in which such an offer,
solicitation or sale would be unlawful prior to registration or qualification
under the securities laws of any such state or other jurisdiction.'
- 'During the six months ended June 30, 2023, we also received net
proceeds of $103 from the sale of shares of our common stock through the Maxim
Sales Agreement.Recent Accounting Pronouncements See Note 2, "Accounting
Policies," to our condensed consolidated financial statements included in this
Quarterly Report on Form 10-Q for a full description of recent accounting
pronouncements.ITEM 3.
QUANTITATIVE AND QUALITATIVE DISCLOSURES ABOUT MARKET
RISK.
Not applicable.ITEM 4.
CONTROLS AND PROCEDURES.
Evaluation of
Disclosure Controls and Procedures Our management (with the participation of
our Principal Executive Officer and Principal Accounting Officer) evaluated
the effectiveness of our disclosure controls and procedures (as defined in
Rules 13a-15(e) and 15d-15(e) under the Exchange Act), as of June 30, 2024.
Disclosure controls and procedures are designed to ensure that information
required to be disclosed by the Company in the reports it files or submits
under the Exchange Act is recorded, processed, summarized and reported on a
timely basis and that such information is accumulated and communicated to
management, including the Principal Executive Officer and the Principal
Accounting Officer, as appropriate, to allow timely decisions regarding
disclosure.
Based on this evaluation, our Principal Executive Officer and
Principal Accounting Officer concluded that these disclosure controls and
procedures are effective.Changes in Internal Control over Financial Reporting
There have been no changes in our internal control over financial reporting
(as defined in Rules 13a-15(f) under the Exchange Act) during the quarter
ended June 30, 2024, that have materially affected, or are reasonably likely
to materially affect, our internal control over financial reporting.'
- 'The Company maintained US Treasury bills with maturities of less than
three months and expects zero credit losses from these securities.
As a
result, the Company did not record an allowance for expected credit losses.Field:
Sequence; Type: Arabic; Name: PageNo 12 Field: /Sequence 5.
EQUITY
TRANSACTIONS IN THE THREE MONTHS ENDED JUNE 30, 2024 2022 At The Market
Offering Agreement with H.C. Wainwright & Co., LLC On March 24, 2022, we
entered into an At The Market Offering Agreement, or the 2022 ATM Agreement,
with H.C. Wainwright & Co., LLC, or Wainwright, which established an at-the-
market equity program pursuant to which we may offer and sell shares of our
common stock from time to time, as set forth in the 2022 ATM Agreement.
The
offering was registered under the Securities Act of 1933, as amended, or the
Securities Act, pursuant to our shelf registration statement on Form S-3
(Registration Statement No. 333-259909), as previously filed with the SEC and
declared effective on October 21, 2021.We filed a prospectus supplement,
dated March 24, 2022, with the SEC that provides for the sale of shares of our
common stock having an aggregate offering price of up to $ 15,000,000 , or the
2022 ATM Shares.Under the 2022 ATM Agreement, Wainwright may sell the 2022
ATM Shares by any method permitted by law and deemed to be an “at the market
offering” as defined in Rule 415 promulgated under the Securities Act,
including sales made directly on the Nasdaq Capital Market, or on any other
existing trading market for the 2022 ATM Shares.
In addition, under the 2022
ATM Agreement, Wainwright may sell the 2022 ATM Shares in privately negotiated
transactions with our consent and in block transactions.
Under certain
circumstances, we may instruct Wainwright not to sell the 2022 ATM Shares if
the sales cannot be effected at or above the price designated by us from time
to time.
We are not obligated to make any sales of the 2022 ATM Shares under
the 2022 ATM Agreement.'
- source_sentence: Hi
sentences:
- 'We have
entered into an Equity Distribution Agreement with Oppenheimer & Co. Inc., or
the Sales Agent, under which we may offer and sell up to $60,000,000 of our
shares of common stock from time to time through our Sales Agent.
Sales of our
shares of common stock, if any, under this prospectus will be made by any
method that is deemed to be an “at-the-market offering” as defined in Rule
415(a)(4) under the Securities Act or, if expressly authorized by us, in
privately negotiated transactions.Each time we wish to issue and sell our
shares of common stock under the Equity Distribution Agreement, we will notify
our Sales Agent of the maximum number of shares to be issued, the dates on
which such sales may be made, any limitation on the number of shares to be
sold in any one day and any minimum price below which sales may not be made.
Once we have instructed our Sales Agent, unless our Sales Agent declines to
accept the terms of such notice, our Sales Agent has agreed to use its
commercially reasonable efforts consistent with its normal trading and sales
practices to sell such shares up to the amount specified on such terms.The
obligations of our Sales Agent under the Equity Distribution Agreement to sell
our shares of common stock are subject to a number of conditions that we must
meet.
The settlement of sales of shares of common stock between us and our
Sales Agent is generally anticipated to occur on the first trading day (unless
we and our Sales Agent have agreed in writing on another date) following the
date on which the sale was made.Sales of our shares of common stock as
contemplated in this prospectus will be settled through the facilities of The
Depository Trust Company or by such other means as we and our Sales Agent may
agree upon.
There is no arrangement for funds to be received in an escrow,
trust or similar arrangement.'
- 'Emerging
Growth Company Status We are an emerging growth company as that term is used
in the Jumpstart Our Business Startups Act of 2012 and, as such, have elected
to comply with certain reduced public company reporting requirements.
Section
107 of the JOBS Act provides that an emerging growth company can take
advantage of the extended transition period provided in Section 7(a)(2)(B) of
the Securities Act for complying with new or revised accounting standards.
In
other words, an emerging growth company can delay the adoption of certain
accounting standards until those standards would otherwise apply to private
companies.
We have elected to take advantage of the benefits of this extended
transition period.
Our financial statements may, therefore, not be comparable
to those of companies that comply with such new or revised accounting
standards.Off-Balance Sheet Arrangements We did not have during the periods
presented, and we do not currently have, any off-balance sheet arrangements,
as defined in the rules and regulations of the Securities and Exchange
Commission.ITEM 3.
QUANTITATIVE AND QUALITATIVE DISCLOSURES ABOUT MARKET RISKWe are a smaller reporting
company as defined by Rule 12b-2 of the Securities
and Exchange Act of 1934, as amended (the “Exchange Act”) and are not required
to provide the information required under this item.ITEM 4.
CONTROLS AND
PROCEDURES Evaluation of Disclosure Controls and Procedures We maintain
“disclosure controls and procedures” as defined in Rules 13a-15(e) and
15d-15(e) under the Securities Exchange Act of 1934, as amended, or the
Exchange Act, that are designed to ensure that information required to be
disclosed in the reports we file and submit under the Exchange Act is
recorded, processed, summarized and reported within the time periods specified
in the SEC’s rules and forms.'
- 'UNITED STATES SECURITIES AND EXCHANGE COMMISSION WASHINGTON, D.C. 20549 FORM
8-K CURRENT REPORT Pursuant to Section 13 or 15(d) of the Securities Exchange
Act of 1934 Date of Report (Date of Earliest Event Reported): August 19, 2024
Federal Home Loan Bank of Pittsburgh (Exact name of registrant as specified in
its charter)Federally Chartered Corporation 000-51395 25-6001324 (State or
other jurisdiction (Commission (I.R.S. Employer of incorporation) File Number)
Identification No.) 601 Grant Street , Pittsburgh , Pennsylvania 15219
(Address of principal executive offices) (Zip Code) Registrant’s telephone
number, including area code: 412 - 288-3400 Not Applicable Former name or
former address, if changed since last report Check the appropriate box below
if the Form 8-K filing is intended to simultaneously satisfy the filing
obligation of the registrant under any of the following provisions:☐ Written
communications pursuant to Rule 425 under the Securities Act (17 CFR 230.425)
☐ Soliciting material pursuant to Rule 14a-12 under the Exchange Act (17 CFR
240.14a-12)☐ Pre-commencement communications pursuant to Rule 14d-2(b) under
the Exchange Act (17 CFR 240.14d-2(b))☐ Pre-commencement communications
pursuant to Rule 13e-4(c) under the Exchange Act (17 CFR 240.13e-4(c))Securities
registered pursuant to Section 12(b) of the Act: Title of each
class Trading Symbol(s) Name of each exchange on which registered — — —
Indicate by check mark whether the registrant is an emerging growth company as
defined in Rule 405 of the Securities Act of 1933 (§230.405 of this chapter)
or Rule 12b-2 of the Securities Exchange Act of 1934 (§240.12b-2 of this
chapter).Emerging growth company ☐ If an emerging growth company, indicate by
check mark if the registrant has elected not to use the extended transition
period for complying with any new or revised financial accounting standards
provided pursuant to Section 13(a) of the Exchange Act.'
- source_sentence: Hi
sentences:
- 'The information
contained herein is intended to be reviewed in its totality, and any
stipulations, conditions or provisos that apply to a given piece of
information in one part of this report should be read as applying mutatis
mutandis to every other instance of such information appearing herein.Item
9.01 Financial Statements and Exhibits.
(d) Exhibits EXHIBIT INDEX Exhibit No.
Description 7.1 (sing_ex71.htm) Letter from Turner.
Stone & Company, L.L.P.
(sing_ex71.htm) 104 Cover Page Interactive Data File (embedded within the
Inline XBRL document.)2 SIGNATURES Pursuant to the requirements of the Stock
Exchange Act of 1934, the registrant has duly caused this report to be signed
on its behalf by the undersigned hereunto duly authorized.SinglePoint Inc.
Dated: August 21, 2024 By: /s/ William Ralston Name: William Ralston Title:
Chief Executive Officer 3'
- 'Open Market
Sale Agreement.
On February 4, 2022, we entered into an Open Market Sale
Agreement with Jefferies LLC, as agent, pursuant to which we may offer and
sell, from time to time, through Jefferies, shares of our common stock having
an aggregate offering price of up to $50,000,000.
On October 12, 2022,
pursuant to this agreement, the Company sold 500,000 shares of common stock in
a single transaction at a price of $10.35 per share generating gross proceeds
of $5.2 million ($4.8 million net of commissions and offering expenses) On
December 1, 2023, pursuant to this agreement, the Company sold 1,034,500
shares of common stock in a single transaction at a price of $14.50 per share,
generating gross proceeds of $15 million ($14.4 million net of commissions and
offering expenses).In April 2024 and May 2024, pursuant to the Open Market
Sale Agreement with Jefferies LLC, as agent, the Company sold 285,714 and
149,700 shares of common stock, respectively, at an average selling price of $
17.55 per share, generating gross proceeds of $7.6 million before deducting
commissions and other offering expenses of $0.3 million.
At June 30, 2024,
$22.2 million of common stock remains available for sale under the Jefferies
agreement.'
- 'On April 18, 2024, the Company
entered into a securities purchase agreement with certain institutional and
accredited investors pursuant to which the Company agreed to sell, in a
registered direct offering, an aggregate of 375,000 shares of its common stock
for gross proceeds of approximately $ 1.2 million under the base prospectus
contained in the 2022 Shelf Registration Statement and a related prospectus
supplement filed with the SEC on April 19, 2024 (the “April 2024 Registered
Direct Offering”).
In a concurrent private placement, the Company also agreed
pursuant to the securities purchase agreement to issue to such investors
warrants to purchase up to 375,000 shares of its common stock at an exercise
price of $ 3.10 per share (the “April 2024 Private Placement”).
The April 2024
Registered Direct Offering and the April 2024 Private Placement closed on
April 19, 2024.
The net proceeds from the offerings, after deducting the
placement agent’s fees and expenses and the Company’s offering expenses, and
excluding the proceeds, if any, from the exercise of the warrants issued in
the offerings, were approximately $ 0.9 million.On April 19, 2024, the
Company determined to increase the number of shares available for sale under
the At The Market Offering Agreement, up to an additional aggregate offering
price of approximately $ 1.1 million, which shares are being offered and sold
pursuant to the 2022 Shelf Registration Statement and a prospectus supplement
and accompanying prospectus filed with the SEC on April 19, 2024 (the
“Subsequent ATM Prospectus Supplement”).As of June 30, 2024, the Company has
offered and sold 334,929 shares of common stock under the Subsequent ATM
Prospectus Supplement for gross proceeds of approximately $ 1.1 million.
The
net proceeds from such offering, after deducting commissions and the Company’s
offering expenses, were approximately $ 1.0 million.'
- source_sentence: Hi
sentences:
- 'Note 9 – Employee Benefit Plans The Company maintains defined
contribution benefit plans under Section 401(k) of the Internal Revenue Code
covering substantially all qualified employees of the Company (the “401(k)
Plan”).
Under the 401(k) Plan, the Company may make discretionary
contributions of up to 100 % of employee contributions.
For the six months
ended June 30, 2024 and 2023, the Company made contributions to the 401(k)
Plan of $ 109,000 and $ 95,000 , respectively.Note 10 – Liquidity The Company
follows “ Presentation of Financial Statements—Going Concern (Subtopic
205-40): Disclosure of Uncertainties about an Entity’s Ability to Continue as
a Going Concern ”.
The Company’s financial statements have been prepared
assuming that it will continue as a going concern, which contemplates
continuity of operations, realization of assets, and liquidation of
liabilities in the normal course of business.
As reflected in the financial
statements, the Company has historically incurred a net loss and has an
accumulated deficit of approximately $ 133,148,000 at June 30, 2024, and net
cash used in operating activities of approximately $ 1,693,000 for the
reporting period then ended.
The Company is implementing its business plan and
generating revenue; however, the Company’s cash position and liquid crypto
assets are sufficient to support its daily operations over the next twelve
months.Our Form S-3 expired on August 14, 2024.
The Company filed a new Form
S-3 on February 14, 2024.
As a result of SEC comments, the new Form S-3 has
not yet gone effective and therefore we may not sell shares under the ATM
Agreement.Note 11 – Subsequent Events The Company evaluates events that have
occurred after the balance sheet date but before the financial statements are
issued.
Based upon the evaluation, the Company did not identify any recognized
or non-recognized subsequent events that would have required adjustment or
disclosure in the financial statements other than disclosed.'
- 'In
connection with his appointment, Mr. Tran entered into the Company’s standard
form of indemnification agreement for its directors, which requires the
Company to, among other things, indemnify its directors against liabilities
that may arise by reason of their status or service.
The agreement also
requires the Company to advance all expenses incurred by directors in
investigating or defending any action, suit or proceeding.
The foregoing
description is qualified in its entirety by the full text of the form of
indemnification agreement, which was filed as Exhibit 10.2 to the Company’s
Current Report on Form 8-K (No. 001-39252) filed on January 12, 2021, and is
incorporated by reference herein.There are no arrangements or understandings
between Mr. Tran and any other persons pursuant to which he was selected as a
director.
Mr. Tran has no family relationships with any of the Company’s
directors or executive officers, and he has no direct or indirect material
interest in any transaction required to be disclosed pursuant to Item 404(a)
of Regulation S-K. Item 9.01.
Financial Statements and Exhibits.
(d) List of
Exhibits Exhibit No.
Description 99.1 Press release dated A
(exhibit991-directorappoint.htm) ugust (exhibit991-directorappoint.htm) 22
(exhibit991-directorappoint.htm) , 2024 (exhibit991-directorappoint.htm) 104
Cover Page Interactive Data File (embedded within the Inline XBRL document)
SIGNATURE Pursuant to the requirements of the Securities Exchange Act of 1934,
the registrant has duly caused this report to be signed on its behalf by the
undersigned thereunto duly authorized.Clover Health Investments, Corp.
Date:
August 22, 2024 By: /s/ Karen M. Soares Name: Karen M. Soares Title: General
Counsel and Corporate Secretary'
- '☐ Item 1.01 Entry into a Material Definitive Agreement.
On
August 21, 2024, Lexaria Bioscience Corp. (the “Company”) entered into a
Capital on Demand™ Sales Agreement (the “Sales Agreement”) with JonesTrading
Institutional Services LLC (the “Agent”), pursuant to which the Company may
issue and sell, from time to time, up to $20,000,000 in aggregate principal
amount of shares (the “Shares”) of the Company’s common stock, par value
$0.001 per share, through or to the Agent, as the Company’s sales agent or
principal.
Any Shares to be offered and sold under the Sales Agreement will be
issued and sold by methods deemed to be an “at-the-market offering” as defined
in Rule 415(a)(4) promulgated under the Securities Act of 1933, as amended
(the “Act”), or in negotiated transactions, if authorized by the Company.
Subject to the terms of the Sales Agreement, the Agent will use reasonable
efforts to sell the Shares from time to time, based upon the Company’s
instructions (including any price, time, or size limits or other customary
parameters or conditions the Company may impose).
The Company cannot provide
any assurances that it will issue any Shares pursuant to the Sales Agreement.The
Company will pay the Agent a commission of 3.0% of the gross sales price
of the Shares sold pursuant to the Sales Agreement, if any.
The Company has
agreed to reimburse the Agent for certain specified expenses as provided in
the Sales Agreement and has also agreed to provide the Agent with customary
indemnification and contribution rights in respect of certain liabilities,
including liabilities under the Act.
The Sales Agreement also contains
customary representations, warranties and covenants.The offering of the
Shares will terminate upon the earliest of (a) the issuance and sale of all of
the Shares by the Agent on the terms and subject to the conditions set forth
in the Sales Agreement or (b) the termination of the Sales Agreement by either
of the parties thereto.'
model-index:
- name: Snowflake-ATM-Avg-v2
results:
- task:
type: custom-triplet
name: Custom Triplet
dataset:
name: dim 768
type: dim_768
metrics:
- type: cosine_accuracy
value: 0.735
name: Cosine Accuracy
- type: dot_accuracy
value: 0.265
name: Dot Accuracy
- type: manhattan_accuracy
value: 0.72
name: Manhattan Accuracy
- type: euclidean_accuracy
value: 0.735
name: Euclidean Accuracy
- type: max_accuracy
value: 0.735
name: Max Accuracy
- task:
type: custom-triplet
name: Custom Triplet
dataset:
name: dim 512
type: dim_512
metrics:
- type: cosine_accuracy
value: 0.735
name: Cosine Accuracy
- type: dot_accuracy
value: 0.265
name: Dot Accuracy
- type: manhattan_accuracy
value: 0.72
name: Manhattan Accuracy
- type: euclidean_accuracy
value: 0.735
name: Euclidean Accuracy
- type: max_accuracy
value: 0.735
name: Max Accuracy
- task:
type: custom-triplet
name: Custom Triplet
dataset:
name: dim 256
type: dim_256
metrics:
- type: cosine_accuracy
value: 0.735
name: Cosine Accuracy
- type: dot_accuracy
value: 0.265
name: Dot Accuracy
- type: manhattan_accuracy
value: 0.72
name: Manhattan Accuracy
- type: euclidean_accuracy
value: 0.735
name: Euclidean Accuracy
- type: max_accuracy
value: 0.735
name: Max Accuracy
- task:
type: custom-triplet
name: Custom Triplet
dataset:
name: dim 128
type: dim_128
metrics:
- type: cosine_accuracy
value: 0.735
name: Cosine Accuracy
- type: dot_accuracy
value: 0.265
name: Dot Accuracy
- type: manhattan_accuracy
value: 0.72
name: Manhattan Accuracy
- type: euclidean_accuracy
value: 0.735
name: Euclidean Accuracy
- type: max_accuracy
value: 0.735
name: Max Accuracy
- task:
type: custom-triplet
name: Custom Triplet
dataset:
name: dim 64
type: dim_64
metrics:
- type: cosine_accuracy
value: 0.735
name: Cosine Accuracy
- type: dot_accuracy
value: 0.265
name: Dot Accuracy
- type: manhattan_accuracy
value: 0.72
name: Manhattan Accuracy
- type: euclidean_accuracy
value: 0.735
name: Euclidean Accuracy
- type: max_accuracy
value: 0.735
name: Max Accuracy
---
# Snowflake-ATM-Avg-v2
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [Snowflake/snowflake-arctic-embed-m-v1.5](https://huggingface.co/Snowflake/snowflake-arctic-embed-m-v1.5). It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [Snowflake/snowflake-arctic-embed-m-v1.5](https://huggingface.co/Snowflake/snowflake-arctic-embed-m-v1.5) <!-- at revision 3b5a16eaf17e47bd997da998988dce5877a57092 -->
- **Maximum Sequence Length:** 512 tokens
- **Output Dimensionality:** 768 tokens
- **Similarity Function:** Cosine Similarity
<!-- - **Training Dataset:** Unknown -->
- **Language:** en
<!-- - **License:** Unknown -->
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("jdaviescmg/Snowflake-ATM-Avg-v2")
# Run inference
sentences = [
'Hi',
'☐ Item 1.01 Entry into a Material Definitive Agreement.\n\nOn\nAugust 21, 2024, Lexaria Bioscience Corp. (the “Company”) entered into a\nCapital on Demand™ Sales Agreement (the “Sales Agreement”) with JonesTrading\nInstitutional Services LLC (the “Agent”), pursuant to which the Company may\nissue and sell, from time to time, up to $20,000,000 in aggregate principal\namount of shares (the “Shares”) of the Company’s common stock, par value\n$0.001 per share, through or to the Agent, as the Company’s sales agent or\nprincipal.\n\nAny Shares to be offered and sold under the Sales Agreement will be\nissued and sold by methods deemed to be an “at-the-market offering” as defined\nin Rule 415(a)(4) promulgated under the Securities Act of 1933, as amended\n(the “Act”), or in negotiated transactions, if authorized by the Company.\n\nSubject to the terms of the Sales Agreement, the Agent will use reasonable\nefforts to sell the Shares from time to time, based upon the Company’s\ninstructions (including any price, time, or size limits or other customary\nparameters or conditions the Company may impose).\n\nThe Company cannot provide\nany assurances that it will issue any Shares pursuant to the Sales Agreement.The Company will pay the Agent a commission of 3.0% of the gross sales price\nof the Shares sold pursuant to the Sales Agreement, if any.\n\nThe Company has\nagreed to reimburse the Agent for certain specified expenses as provided in\nthe Sales Agreement and has also agreed to provide the Agent with customary\nindemnification and contribution rights in respect of certain liabilities,\nincluding liabilities under the Act.\n\nThe Sales Agreement also contains\ncustomary representations, warranties and covenants.The offering of the\nShares will terminate upon the earliest of (a) the issuance and sale of all of\nthe Shares by the Agent on the terms and subject to the conditions set forth\nin the Sales Agreement or (b) the termination of the Sales Agreement by either\nof the parties thereto.',
'Note 9 – Employee Benefit Plans The Company maintains defined\ncontribution benefit plans under Section 401(k) of the Internal Revenue Code\ncovering substantially all qualified employees of the Company (the “401(k)\nPlan”).\n\nUnder the 401(k) Plan, the Company may make discretionary\ncontributions of up to 100 % of employee contributions.\n\nFor the six months\nended June 30, 2024 and 2023, the Company made contributions to the 401(k)\nPlan of $ 109,000 and $ 95,000 , respectively.Note 10 – Liquidity The Company\nfollows “ Presentation of Financial Statements—Going Concern (Subtopic\n205-40): Disclosure of Uncertainties about an Entity’s Ability to Continue as\na Going Concern ”.\n\nThe Company’s financial statements have been prepared\nassuming that it will continue as a going concern, which contemplates\ncontinuity of operations, realization of assets, and liquidation of\nliabilities in the normal course of business.\n\nAs reflected in the financial\nstatements, the Company has historically incurred a net loss and has an\naccumulated deficit of approximately $ 133,148,000 at June 30, 2024, and net\ncash used in operating activities of approximately $ 1,693,000 for the\nreporting period then ended.\n\nThe Company is implementing its business plan and\ngenerating revenue; however, the Company’s cash position and liquid crypto\nassets are sufficient to support its daily operations over the next twelve\nmonths.Our Form S-3 expired on August 14, 2024.\n\nThe Company filed a new Form\nS-3 on February 14, 2024.\n\nAs a result of SEC comments, the new Form S-3 has\nnot yet gone effective and therefore we may not sell shares under the ATM\nAgreement.Note 11 – Subsequent Events The Company evaluates events that have\noccurred after the balance sheet date but before the financial statements are\nissued.\n\nBased upon the evaluation, the Company did not identify any recognized\nor non-recognized subsequent events that would have required adjustment or\ndisclosure in the financial statements other than disclosed.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
## Evaluation
### Metrics
#### Custom Triplet
* Dataset: `dim_768`
* Evaluated with <code>__main__.CustomTripletEvaluator</code>
| Metric | Value |
|:--------------------|:----------|
| **cosine_accuracy** | **0.735** |
| dot_accuracy | 0.265 |
| manhattan_accuracy | 0.72 |
| euclidean_accuracy | 0.735 |
| max_accuracy | 0.735 |
#### Custom Triplet
* Dataset: `dim_512`
* Evaluated with <code>__main__.CustomTripletEvaluator</code>
| Metric | Value |
|:--------------------|:----------|
| **cosine_accuracy** | **0.735** |
| dot_accuracy | 0.265 |
| manhattan_accuracy | 0.72 |
| euclidean_accuracy | 0.735 |
| max_accuracy | 0.735 |
#### Custom Triplet
* Dataset: `dim_256`
* Evaluated with <code>__main__.CustomTripletEvaluator</code>
| Metric | Value |
|:--------------------|:----------|
| **cosine_accuracy** | **0.735** |
| dot_accuracy | 0.265 |
| manhattan_accuracy | 0.72 |
| euclidean_accuracy | 0.735 |
| max_accuracy | 0.735 |
#### Custom Triplet
* Dataset: `dim_128`
* Evaluated with <code>__main__.CustomTripletEvaluator</code>
| Metric | Value |
|:--------------------|:----------|
| **cosine_accuracy** | **0.735** |
| dot_accuracy | 0.265 |
| manhattan_accuracy | 0.72 |
| euclidean_accuracy | 0.735 |
| max_accuracy | 0.735 |
#### Custom Triplet
* Dataset: `dim_64`
* Evaluated with <code>__main__.CustomTripletEvaluator</code>
| Metric | Value |
|:--------------------|:----------|
| **cosine_accuracy** | **0.735** |
| dot_accuracy | 0.265 |
| manhattan_accuracy | 0.72 |
| euclidean_accuracy | 0.735 |
| max_accuracy | 0.735 |
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### Unnamed Dataset
* Size: 800 training samples
* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
* Approximate statistics based on the first 1000 samples:
| | sentence1 | sentence2 | label |
|:--------|:-------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|:------------------------------------------------|
| type | string | string | int |
| details | <ul><li>min: 3 tokens</li><li>mean: 3.0 tokens</li><li>max: 3 tokens</li></ul> | <ul><li>min: 35 tokens</li><li>mean: 371.57 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>0: ~50.00%</li><li>1: ~50.00%</li></ul> |
* Samples:
| sentence1 | sentence2 | label |
|:----------------|:------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------|
| <code>Hi</code> | <code>8. COMMON STOCK [a] Authorized 150,000,000 authorized<br>common shares, par value of $ 0.001 , and 5,000,000 preferred shares, par<br>value of $ 0.001 .<br><br>[b] Issued and outstanding shares At-the-Market Sales<br>AgreementOn December 21, 2021, we entered into an At-the-Market Offering<br>Sales Agreement, or ATM, with Virtu Americas, LLC, as sales agent.<br><br>The ATM was<br>terminated on February 29, 2024, and no further sales of our common stock will<br>be made pursuant to the ATM.<br><br>Since entry into the ATM, through the date of<br>termination of the ATM, we offered and sold an aggregate of 200,000 shares of<br>our common stock.<br><br>These aggregate sales resulted in gross proceeds to us of<br>approximately $ 1.5 million.<br><br>During the three and six months ended June 30,<br>2024, we did no t sell any shares of our common stock pursuant to the ATM.May<br>2023 Registered Direct Offering In May 2023, we entered into a securities<br>purchase agreement with certain purchasers, pursuant to which we sold<br>3,000,000 shares of common stock at a price of $ 5.50 per share in a<br>registered direct offering.<br><br>The offering of the shares was made pursuant to<br>our shelf registration statement on Form S-3 including the prospectus dated<br>January 5, 2022 contained therein, and the prospectus supplement dated May 25,<br>2023. We received approximately $ 15.3 million in net proceeds from the<br>registered direct offering after deducting placement agent fees and offering<br>expenses.February 2024 Registered Direct Offering and Concurrent Private<br>PlacementIn February 2024, we entered into a securities purchase agreement<br>with certain purchasers, pursuant to which we sold 13,086,151 shares of common<br>stock at a price of $ 4.585 per share in a registered direct offering.<br><br>The<br>offering of the shares was made pursuant to our shelf registration statement<br>on Form S-3, including the prospectus dated January 5, 2022 contained therein,<br>and the prospectus supplement dated February 28, 2024.</code> | <code>1</code> |
| <code>Hi</code> | <code>The foregoing description of the Note does not purport to be complete and is<br>subject to, and is qualified in its entirety by reference to, the full text of<br>the Note, which is attached as Exhibit 10.1 to this Current Report on Form<br>8-K, and is incorporated herein by reference.Item 2.03.<br><br>Creation of a Direct<br>Financial Obligation or an Obligation under an Off-Balance Sheet Arrangement<br>of a Registrant.<br><br>The disclosure provided in Item 1.01 of this Current Report<br>on Form 8-K is hereby incorporated by reference into this Item 2.03.Item<br>8.01.Other Events.<br><br>The Company is supplementing the Company’s risk factors in<br>its Annual Report on Form 10-K filed with the SEC on March 29, 2024, and<br>Quarterly Reports on Form 10-Q for the quarters ended March 31, 2024 and June<br>30, 2024, filed with the SEC on May 10, 2024 and August 14, 2024,<br>respectively, with the risk factor set forth below.Servicing our debt will<br>require a significant amount of cash, and we may not have sufficient cash flow<br>from our business to pay our debt.<br><br>Our ability to make scheduled payments of<br>the principal of, to pay interest on or to refinance our indebtedness depends<br>on our future performance, which is subject to economic, financial,<br>competitive and other factors beyond our control.<br><br>We had, as of June 30, 2024,<br>approximately (i) $16.1 million in working capital, (ii) $2.4 million in cash<br>and cash equivalents, and (iii) $13.6 million of outstanding indebtedness, net<br>of discounts.<br><br>In addition, on August 15, 2024, we amended and restated the<br>unsecured promissory note and guaranty previously issued to JXVII Trust that<br>increased the principal amount from $7.6 million to $10.0 million.</code> | <code>0</code> |
| <code>Hi</code> | <code>The Company<br>incurred costs of approximately $0.9 million related to the execution of the<br>Purchase Agreement.<br><br>Of the total costs incurred, approximately $0.6 million<br>was paid in Common Stock to Lincoln Park as a commitment fee and $ 0.03<br>million to reimburse Lincoln Park for expenses.<br><br>These transaction costs were<br>included in other income / (expenses), net in the consolidated statement of<br>operations.<br><br>Approximately $ 0.2 million was incurred for legal fees, which<br>were included in administrative and selling expenses on the consolidated<br>statement of operations.During the year ended December 31, 2023, the Company<br>issued and sold an aggregate of 293,509 shares pursuant to the Purchase<br>Agreement and received net proceeds of $ 5.5 million.During the year ended<br>December 31, 2023, the Company incurred approximately $ 0.3 million of<br>expenses, related to the discount on the issuance of common stock to Lincoln<br>Park, which is included in other income / (expenses), net in the consolidated<br>statement of operations.<br><br>As the Company’s common stock price is below $15.00<br>per share, the Company is unable to utilize the facility.At the Market<br>Offering Agreement On June 2, 2023, the Company entered into an At The Market<br>Offering Agreement (the “ATM Agreement”) with H.C. Wainwright & Co., LLC, as<br>sales agent (the “Agent”), to create an at-the-market equity program under<br>which it may sell up to $50 million of shares of the Company’s common stock<br>(the “Shares”) from time to time through the Agent (the “ATM Offering”).<br><br>Under<br>the ATM Agreement, the Agent will be entitled to a commission at a fixed rate<br>of 3.0 % of the gross proceeds from each sale of Shares under the ATM<br>Agreement.</code> | <code>1</code> |
* Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
```json
{
"loss": "CustomContrastiveLoss",
"matryoshka_dims": [
768,
512,
256,
128,
64
],
"matryoshka_weights": [
1,
1,
1,
1,
1
],
"n_dims_per_step": -1
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: epoch
- `per_device_train_batch_size`: 32
- `per_device_eval_batch_size`: 16
- `gradient_accumulation_steps`: 16
- `learning_rate`: 1e-05
- `num_train_epochs`: 10
- `warmup_ratio`: 0.1
- `use_mps_device`: True
- `optim`: adamw_hf
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: epoch
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 32
- `per_device_eval_batch_size`: 16
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 16
- `eval_accumulation_steps`: None
- `learning_rate`: 1e-05
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1.0
- `num_train_epochs`: 10
- `max_steps`: -1
- `lr_scheduler_type`: linear
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.1
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: True
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: False
- `fp16`: False
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: None
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: False
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_hf
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: False
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `dispatch_batches`: None
- `split_batches`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `batch_sampler`: batch_sampler
- `multi_dataset_batch_sampler`: proportional
</details>
### Training Logs
| Epoch | Step | Training Loss | dim_128_cosine_accuracy | dim_256_cosine_accuracy | dim_512_cosine_accuracy | dim_64_cosine_accuracy | dim_768_cosine_accuracy |
|:-----:|:----:|:-------------:|:-----------------------:|:-----------------------:|:-----------------------:|:----------------------:|:-----------------------:|
| 0.64 | 1 | - | 0.7 | 0.7 | 0.7 | 0.7 | 0.7 |
| 1.92 | 3 | - | 0.72 | 0.72 | 0.72 | 0.72 | 0.72 |
| 2.56 | 4 | - | 0.72 | 0.72 | 0.72 | 0.72 | 0.72 |
| 3.84 | 6 | - | 0.71 | 0.71 | 0.71 | 0.71 | 0.71 |
| 4.48 | 7 | - | 0.715 | 0.715 | 0.715 | 0.715 | 0.715 |
| 5.76 | 9 | - | 0.71 | 0.71 | 0.71 | 0.71 | 0.71 |
| 6.4 | 10 | 0.1627 | 0.715 | 0.715 | 0.715 | 0.715 | 0.715 |
| 0.64 | 1 | - | 0.715 | 0.715 | 0.715 | 0.715 | 0.715 |
| 1.92 | 3 | - | 0.72 | 0.72 | 0.72 | 0.72 | 0.72 |
| 2.56 | 4 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 3.84 | 6 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 4.48 | 7 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 5.76 | 9 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 6.4 | 10 | 0.0739 | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 0.64 | 1 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 1.92 | 3 | - | 0.735 | 0.735 | 0.735 | 0.735 | 0.735 |
| 2.56 | 4 | - | 0.735 | 0.735 | 0.735 | 0.735 | 0.735 |
| 3.84 | 6 | - | 0.74 | 0.74 | 0.74 | 0.74 | 0.74 |
| 4.48 | 7 | - | 0.74 | 0.74 | 0.74 | 0.74 | 0.74 |
| 5.76 | 9 | - | 0.74 | 0.74 | 0.74 | 0.74 | 0.74 |
| 6.4 | 10 | 0.0521 | 0.74 | 0.74 | 0.74 | 0.74 | 0.74 |
| 0.64 | 1 | - | 0.74 | 0.74 | 0.74 | 0.74 | 0.74 |
| 1.92 | 3 | - | 0.68 | 0.68 | 0.68 | 0.68 | 0.68 |
| 2.56 | 4 | - | 0.665 | 0.665 | 0.665 | 0.665 | 0.665 |
| 3.84 | 6 | - | 0.695 | 0.695 | 0.695 | 0.695 | 0.695 |
| 4.48 | 7 | - | 0.715 | 0.715 | 0.715 | 0.715 | 0.715 |
| 5.76 | 9 | - | 0.72 | 0.72 | 0.72 | 0.72 | 0.72 |
| 6.4 | 10 | 0.249 | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 0.64 | 1 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 1.92 | 3 | - | 0.735 | 0.735 | 0.735 | 0.735 | 0.735 |
| 2.56 | 4 | - | 0.735 | 0.735 | 0.735 | 0.735 | 0.735 |
| 3.84 | 6 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 4.48 | 7 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 5.76 | 9 | - | 0.72 | 0.72 | 0.72 | 0.72 | 0.72 |
| 6.4 | 10 | 0.0244 | 0.72 | 0.72 | 0.72 | 0.72 | 0.72 |
| 0.64 | 1 | - | 0.72 | 0.72 | 0.72 | 0.72 | 0.72 |
| 1.92 | 3 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 2.56 | 4 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 3.84 | 6 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 4.48 | 7 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 5.12 | 8 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 0.64 | 1 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 1.92 | 3 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 2.56 | 4 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 3.84 | 6 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 4.48 | 7 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 5.76 | 9 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 6.4 | 10 | 0.0123 | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 7.68 | 12 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 0.64 | 1 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 1.92 | 3 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 2.56 | 4 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 3.84 | 6 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 4.48 | 7 | - | 0.725 | 0.725 | 0.725 | 0.725 | 0.725 |
| 5.76 | 9 | - | 0.73 | 0.73 | 0.73 | 0.73 | 0.73 |
| 6.4 | 10 | 0.0078 | 0.735 | 0.735 | 0.735 | 0.735 | 0.735 |
### Framework Versions
- Python: 3.12.5
- Sentence Transformers: 3.0.1
- Transformers: 4.41.2
- PyTorch: 2.4.1
- Accelerate: 0.34.2
- Datasets: 2.19.1
- Tokenizers: 0.19.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### MatryoshkaLoss
```bibtex
@misc{kusupati2024matryoshka,
title={Matryoshka Representation Learning},
author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
year={2024},
eprint={2205.13147},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
AhmedSSoliman/DistilBERT-Marian-Model-on-DJANGO | AhmedSSoliman | translation | [
"transformers",
"pytorch",
"encoder-decoder",
"text2text-generation",
"Code Generation",
"Machine translation",
"Text generation",
"translation",
"en",
"dataset:AhmedSSoliman/DJANGO",
"license:mit",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,673,474,083,000 | 2023-07-30T12:01:43 | 13 | 0 | ---
datasets:
- AhmedSSoliman/DJANGO
language:
- en
license: mit
metrics:
- bleu
- accuracy
pipeline_tag: translation
tags:
- Code Generation
- Machine translation
- Text generation
---
| [
"TRANSLATION"
] | Non_BioNLP |
SoyGema/english-guyarati | SoyGema | translation | [
"transformers",
"pytorch",
"safetensors",
"t5",
"text2text-generation",
"generated_from_trainer",
"translation",
"en",
"gu",
"dataset:opus100",
"base_model:google-t5/t5-small",
"base_model:finetune:google-t5/t5-small",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,694,256,593,000 | 2023-09-11T06:51:34 | 16 | 0 | ---
base_model: t5-small
datasets:
- opus100
language:
- en
- gu
license: apache-2.0
metrics:
- bleu
pipeline_tag: translation
tags:
- generated_from_trainer
model-index:
- name: english-guyarati
results:
- task:
type: translation
name: Translation
dataset:
name: opus100 en-gu
type: opus100
config: en-gu
split: validation
args: en-gu
metrics:
- type: bleu
value: 48.3798
name: Bleu
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# english-guyarati
This model is a fine-tuned version of [t5-small](https://huggingface.co/t5-small) on the opus100 en-gu dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3595
- Bleu: 48.3798
- Gen Len: 14.5
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3.0
### Training results
### Framework versions
- Transformers 4.32.0.dev0
- Pytorch 2.0.1
- Datasets 2.14.4
- Tokenizers 0.13.3 | [
"TRANSLATION"
] | Non_BioNLP |
google/t5-efficient-small-el4 | google | text2text-generation | [
"transformers",
"pytorch",
"tf",
"jax",
"t5",
"text2text-generation",
"deep-narrow",
"en",
"dataset:c4",
"arxiv:2109.10686",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"region:us"
] | 1,646,263,745,000 | 2023-01-24T16:49:01 | 118 | 0 | ---
datasets:
- c4
language:
- en
license: apache-2.0
tags:
- deep-narrow
inference: false
---
# T5-Efficient-SMALL-EL4 (Deep-Narrow version)
T5-Efficient-SMALL-EL4 is a variation of [Google's original T5](https://ai.googleblog.com/2020/02/exploring-transfer-learning-with-t5.html) following the [T5 model architecture](https://huggingface.co/docs/transformers/model_doc/t5).
It is a *pretrained-only* checkpoint and was released with the
paper **[Scale Efficiently: Insights from Pre-training and Fine-tuning Transformers](https://arxiv.org/abs/2109.10686)**
by *Yi Tay, Mostafa Dehghani, Jinfeng Rao, William Fedus, Samira Abnar, Hyung Won Chung, Sharan Narang, Dani Yogatama, Ashish Vaswani, Donald Metzler*.
In a nutshell, the paper indicates that a **Deep-Narrow** model architecture is favorable for **downstream** performance compared to other model architectures
of similar parameter count.
To quote the paper:
> We generally recommend a DeepNarrow strategy where the model’s depth is preferentially increased
> before considering any other forms of uniform scaling across other dimensions. This is largely due to
> how much depth influences the Pareto-frontier as shown in earlier sections of the paper. Specifically, a
> tall small (deep and narrow) model is generally more efficient compared to the base model. Likewise,
> a tall base model might also generally more efficient compared to a large model. We generally find
> that, regardless of size, even if absolute performance might increase as we continue to stack layers,
> the relative gain of Pareto-efficiency diminishes as we increase the layers, converging at 32 to 36
> layers. Finally, we note that our notion of efficiency here relates to any one compute dimension, i.e.,
> params, FLOPs or throughput (speed). We report all three key efficiency metrics (number of params,
> FLOPS and speed) and leave this decision to the practitioner to decide which compute dimension to
> consider.
To be more precise, *model depth* is defined as the number of transformer blocks that are stacked sequentially.
A sequence of word embeddings is therefore processed sequentially by each transformer block.
## Details model architecture
This model checkpoint - **t5-efficient-small-el4** - is of model type **Small** with the following variations:
- **el** is **4**
It has **54.23** million parameters and thus requires *ca.* **216.9 MB** of memory in full precision (*fp32*)
or **108.45 MB** of memory in half precision (*fp16* or *bf16*).
A summary of the *original* T5 model architectures can be seen here:
| Model | nl (el/dl) | ff | dm | kv | nh | #Params|
| ----| ---- | ---- | ---- | ---- | ---- | ----|
| Tiny | 4/4 | 1024 | 256 | 32 | 4 | 16M|
| Mini | 4/4 | 1536 | 384 | 32 | 8 | 31M|
| Small | 6/6 | 2048 | 512 | 32 | 8 | 60M|
| Base | 12/12 | 3072 | 768 | 64 | 12 | 220M|
| Large | 24/24 | 4096 | 1024 | 64 | 16 | 738M|
| Xl | 24/24 | 16384 | 1024 | 128 | 32 | 3B|
| XXl | 24/24 | 65536 | 1024 | 128 | 128 | 11B|
whereas the following abbreviations are used:
| Abbreviation | Definition |
| ----| ---- |
| nl | Number of transformer blocks (depth) |
| dm | Dimension of embedding vector (output vector of transformers block) |
| kv | Dimension of key/value projection matrix |
| nh | Number of attention heads |
| ff | Dimension of intermediate vector within transformer block (size of feed-forward projection matrix) |
| el | Number of transformer blocks in the encoder (encoder depth) |
| dl | Number of transformer blocks in the decoder (decoder depth) |
| sh | Signifies that attention heads are shared |
| skv | Signifies that key-values projection matrices are tied |
If a model checkpoint has no specific, *el* or *dl* than both the number of encoder- and decoder layers correspond to *nl*.
## Pre-Training
The checkpoint was pretrained on the [Colossal, Cleaned version of Common Crawl (C4)](https://huggingface.co/datasets/c4) for 524288 steps using
the span-based masked language modeling (MLM) objective.
## Fine-Tuning
**Note**: This model is a **pretrained** checkpoint and has to be fine-tuned for practical usage.
The checkpoint was pretrained in English and is therefore only useful for English NLP tasks.
You can follow on of the following examples on how to fine-tune the model:
*PyTorch*:
- [Summarization](https://github.com/huggingface/transformers/tree/master/examples/pytorch/summarization)
- [Question Answering](https://github.com/huggingface/transformers/blob/master/examples/pytorch/question-answering/run_seq2seq_qa.py)
- [Text Classification](https://github.com/huggingface/transformers/tree/master/examples/pytorch/text-classification) - *Note*: You will have to slightly adapt the training example here to make it work with an encoder-decoder model.
*Tensorflow*:
- [Summarization](https://github.com/huggingface/transformers/tree/master/examples/tensorflow/summarization)
- [Text Classification](https://github.com/huggingface/transformers/tree/master/examples/tensorflow/text-classification) - *Note*: You will have to slightly adapt the training example here to make it work with an encoder-decoder model.
*JAX/Flax*:
- [Summarization](https://github.com/huggingface/transformers/tree/master/examples/flax/summarization)
- [Text Classification](https://github.com/huggingface/transformers/tree/master/examples/flax/text-classification) - *Note*: You will have to slightly adapt the training example here to make it work with an encoder-decoder model.
## Downstream Performance
TODO: Add table if available
## Computational Complexity
TODO: Add table if available
## More information
We strongly recommend the reader to go carefully through the original paper **[Scale Efficiently: Insights from Pre-training and Fine-tuning Transformers](https://arxiv.org/abs/2109.10686)** to get a more nuanced understanding of this model checkpoint.
As explained in the following [issue](https://github.com/google-research/google-research/issues/986#issuecomment-1035051145), checkpoints including the *sh* or *skv*
model architecture variations have *not* been ported to Transformers as they are probably of limited practical usage and are lacking a more detailed description. Those checkpoints are kept [here](https://huggingface.co/NewT5SharedHeadsSharedKeyValues) as they might be ported potentially in the future. | [
"TEXT_CLASSIFICATION",
"QUESTION_ANSWERING",
"SUMMARIZATION"
] | Non_BioNLP |
varun-v-rao/bart-base-lora-885K-snli-model1 | varun-v-rao | text-classification | [
"transformers",
"tensorboard",
"safetensors",
"bart",
"text-classification",
"generated_from_trainer",
"dataset:stanfordnlp/snli",
"base_model:facebook/bart-base",
"base_model:finetune:facebook/bart-base",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,718,822,037,000 | 2024-06-19T22:49:03 | 4 | 0 | ---
base_model: facebook/bart-base
datasets:
- stanfordnlp/snli
license: apache-2.0
metrics:
- accuracy
tags:
- generated_from_trainer
model-index:
- name: bart-base-lora-885K-snli-model1
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: snli
type: stanfordnlp/snli
metrics:
- type: accuracy
value: 0.8271692745376956
name: Accuracy
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# bart-base-lora-885K-snli-model1
This model is a fine-tuned version of [facebook/bart-base](https://huggingface.co/facebook/bart-base) on the snli dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4486
- Accuracy: 0.8272
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 256
- eval_batch_size: 128
- seed: 30
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------:|
| 0.6339 | 1.0 | 2146 | 0.5079 | 0.7996 |
| 0.5725 | 2.0 | 4292 | 0.4618 | 0.8215 |
| 0.5537 | 3.0 | 6438 | 0.4486 | 0.8272 |
### Framework versions
- Transformers 4.35.2
- Pytorch 2.1.1+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
Yongxin-Guo/VTG-LLM | Yongxin-Guo | null | [
"dense-video-caption",
"video-highlight-detection",
"video-summarization",
"moment-retrieval",
"dataset:Yongxin-Guo/VTG-IT",
"arxiv:2405.13382",
"license:apache-2.0",
"region:us"
] | 1,716,262,231,000 | 2024-06-19T08:29:04 | 0 | 3 | ---
datasets:
- Yongxin-Guo/VTG-IT
license: apache-2.0
tags:
- dense-video-caption
- video-highlight-detection
- video-summarization
- moment-retrieval
---
[VTG-LLM: Integrating Timestamp Knowledge into Video LLMs for Enhanced Video Temporal Grounding](https://arxiv.org/abs/2405.13382)
## Overview
We introduce
- VTG-IT-120K, a high-quality and comprehensive instruction tuning dataset that covers VTG tasks such as moment retrieval (63.2K), dense video captioning (37.2K), video summarization (15.2K), and video highlight detection (3.9K).
- VTG-LLM, which (1) effectively integrates timestamp knowledge into visual tokens; (2) incorporates absolute-time tokens that specifically handle timestamp knowledge, thereby avoiding concept shifts; and (3) introduces a lightweight, high-performance slot-based token compression method to facilitate the sampling of more video frames.
## How to Use
Please refer to [GitHub repo](https://github.com/gyxxyg/VTG-LLM) for details.
## Citation
If you find this repository helpful for your project, please consider citing:
```
@article{guo2024vtg,
title={VTG-LLM: Integrating Timestamp Knowledge into Video LLMs for Enhanced Video Temporal Grounding},
author={Guo, Yongxin and Liu, Jingyu and Li, Mingda and Tang, Xiaoying and Chen, Xi and Zhao, Bo},
journal={arXiv preprint arXiv:2405.13382},
year={2024}
}
``` | [
"SUMMARIZATION"
] | Non_BioNLP |
mrapacz/interlinear-en-greta-emb-sum-normalized-bh | mrapacz | text2text-generation | [
"transformers",
"pytorch",
"morph-t5-sum",
"text2text-generation",
"en",
"dataset:mrapacz/greek-interlinear-translations",
"license:cc-by-sa-4.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,739,017,736,000 | 2025-02-21T21:31:34 | 16 | 0 | ---
base_model:
- GreTa
datasets:
- mrapacz/greek-interlinear-translations
language:
- en
library_name: transformers
license: cc-by-sa-4.0
metrics:
- bleu
---
# Model Card for Ancient Greek to English Interlinear Translation Model
This model performs interlinear translation from Ancient Greek to English, maintaining word-level alignment between source and target texts.
You can find the source code used for training this and other models trained as part of this project in the [GitHub repository](https://github.com/mrapacz/loreslm-interlinear-translation).
## Model Details
### Model Description
- **Developed By:** Maciej Rapacz, AGH University of Kraków
- **Model Type:** MorphT5SumForConditionalGeneration
- **Base Model:** GreTa
- **Tokenizer:** GreTa
- **Language(s):** Ancient Greek (source) → English (target)
- **License:** CC BY-NC-SA 4.0
- **Tag Set:** BH (Bible Hub)
- **Text Preprocessing:** Normalized
- **Morphological Encoding:** emb-sum
### Model Performance
- **BLEU Score:** 51.93
- **SemScore:** 0.84
### Model Sources
- **Repository:** https://github.com/mrapacz/loreslm-interlinear-translation
- **Paper:** https://aclanthology.org/2025.loreslm-1.11/
## Usage Example
> **Note**: This model uses a modification of T5-family models that includes dedicated embedding layers for encoding morphological information. To load these models, install the [morpht5](https://github.com/mrapacz/loreslm-interlinear-translation/blob/master/morpht5/README.md) package:
> ```bash
> pip install morpht5
> ```
```python
>>> from morpht5 import MorphT5SumForConditionalGeneration, MorphT5Tokenizer
>>> text = ['λεγει', 'αυτω', 'ο', 'ιησους', 'εγειρε', 'αρον', 'τον', 'κραβαττον', 'σου', 'και', 'περιπατει']
>>> tags = ['V-PIA-3S', 'PPro-DM3S', 'Art-NMS', 'N-NMS', 'V-PMA-2S', 'V-AMA-2S', 'Art-AMS', 'N-AMS', 'PPro-G2S', 'Conj', 'V-PMA-2S']
>>> tokenizer = MorphT5Tokenizer.from_pretrained("mrapacz/interlinear-en-greta-emb-sum-normalized-bh")
>>> inputs = tokenizer(
text=text,
morph_tags=tags,
return_tensors="pt"
)
>>> model = MorphT5SumForConditionalGeneration.from_pretrained("mrapacz/interlinear-en-greta-emb-sum-normalized-bh")
>>> outputs = model.generate(
**inputs,
max_new_tokens=100,
early_stopping=True,
)
>>> decoded = tokenizer.decode(outputs[0], skip_special_tokens=True, keep_block_separator=True)
>>> decoded = decoded.replace(tokenizer.target_block_separator_token, " | ")
>>> decoded
'says | to him | - | jesus | arise | take up | the | mat | of you | and | walk'
```
## Citation
If you use this model, please cite the following paper:
```
@inproceedings{rapacz-smywinski-pohl-2025-low,
title = "Low-Resource Interlinear Translation: Morphology-Enhanced Neural Models for {A}ncient {G}reek",
author = "Rapacz, Maciej and
Smywi{\'n}ski-Pohl, Aleksander",
editor = "Hettiarachchi, Hansi and
Ranasinghe, Tharindu and
Rayson, Paul and
Mitkov, Ruslan and
Gaber, Mohamed and
Premasiri, Damith and
Tan, Fiona Anting and
Uyangodage, Lasitha",
booktitle = "Proceedings of the First Workshop on Language Models for Low-Resource Languages",
month = jan,
year = "2025",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.loreslm-1.11/",
pages = "145--165",
abstract = "Contemporary machine translation systems prioritize fluent, natural-sounding output with flexible word ordering. In contrast, interlinear translation maintains the source text`s syntactic structure by aligning target language words directly beneath their source counterparts. Despite its importance in classical scholarship, automated approaches to interlinear translation remain understudied. We evaluated neural interlinear translation from Ancient Greek to English and Polish using four transformer-based models: two Ancient Greek-specialized (GreTa and PhilTa) and two general-purpose multilingual models (mT5-base and mT5-large). Our approach introduces novel morphological embedding layers and evaluates text preprocessing and tag set selection across 144 experimental configurations using a word-aligned parallel corpus of the Greek New Testament. Results show that morphological features through dedicated embedding layers significantly enhance translation quality, improving BLEU scores by 35{\%} (44.67 {\textrightarrow} 60.40) for English and 38{\%} (42.92 {\textrightarrow} 59.33) for Polish compared to baseline models. PhilTa achieves state-of-the-art performance for English, while mT5-large does so for Polish. Notably, PhilTa maintains stable performance using only 10{\%} of training data. Our findings challenge the assumption that modern neural architectures cannot benefit from explicit morphological annotations. While preprocessing strategies and tag set selection show minimal impact, the substantial gains from morphological embeddings demonstrate their value in low-resource scenarios."
}
``` | [
"TRANSLATION"
] | Non_BioNLP |
nianlong/memsum-arxiv-summarization | nianlong | null | [
"license:apache-2.0",
"region:us"
] | 1,690,472,626,000 | 2024-03-29T16:00:23 | 0 | 2 | ---
license: apache-2.0
---
[](http://dx.doi.org/10.18653/v1/2022.acl-long.450)
# MemSum: Extractive Summarization of Long Documents Using Multi-Step Episodic Markov Decision Processes
Code for ACL 2022 paper on the topic of long document extractive summarization: [MemSum: Extractive Summarization of Long Documents Using Multi-Step Episodic Markov Decision Processes](https://aclanthology.org/2022.acl-long.450/).
## Set Up Environment
1. create an Anaconda environment, with a name e.g. memsum
**Note**: Without further notification, the following commands need to be run in the working directory where this jupyter notebook is located.
```bash
conda create -n memsum python=3.10
```
2. activate this environment
```bash
source activate memsum
```
3. Install pytorch (GPU version).
```bash
pip install torch torchvision torchaudio
```
4. Install dependencies via pip
```bash
pip install -r requirements.txt
```
## Download Datasets and Pretrained Model Checkpoints
### Download All Datasets Used in the Paper
```python
import os
import subprocess
import wget
for dataset_name in [ "arxiv", "pubmed", "gov-report"]:
print(dataset_name)
os.makedirs( "data/"+dataset_name, exist_ok=True )
## dataset is stored at huggingface hub
train_dataset_path = f"https://huggingface.co/datasets/nianlong/long-doc-extractive-summarization-{dataset_name}/resolve/main/train.jsonl"
val_dataset_path = f"https://huggingface.co/datasets/nianlong/long-doc-extractive-summarization-{dataset_name}/resolve/main/val.jsonl"
test_dataset_path = f"https://huggingface.co/datasets/nianlong/long-doc-extractive-summarization-{dataset_name}/resolve/main/test.jsonl"
wget.download( train_dataset_path, out = "data/"+dataset_name )
wget.download( val_dataset_path, out = "data/"+dataset_name )
wget.download( test_dataset_path, out = "data/"+dataset_name )
```
### Download Pretrained Model Checkpoints
The trained MemSum model checkpoints are stored on huggingface hub
```python
from huggingface_hub import snapshot_download
## download the pretrained glove word embedding (200 dimension)
snapshot_download('nianlong/memsum-word-embedding', local_dir = "model/word_embedding" )
## download model checkpoint on the arXiv dataset
snapshot_download('nianlong/memsum-arxiv-summarization', local_dir = "model/memsum-arxiv" )
## download model checkpoint on the PubMed dataset
snapshot_download('nianlong/memsum-pubmed-summarization', local_dir = "model/memsum-pubmed" )
## download model checkpoint on the Gov-Report dataset
snapshot_download('nianlong/memsum-gov-report-summarization', local_dir = "model/memsum-gov-report" )
```
## Testing Pretrained Model on a Given Dataset
For example, the following command test the performance of the full MemSum model. Berfore runing these codes, make sure current working directory is the main directory "MemSum/" where the .py file summarizers.py is located.
```python
from src.summarizer import MemSum
from tqdm import tqdm
from rouge_score import rouge_scorer
import json
import numpy as np
```
```python
rouge_cal = rouge_scorer.RougeScorer(['rouge1','rouge2', 'rougeLsum'], use_stemmer=True)
memsum_arxiv = MemSum( "model/memsum-arxiv/model.pt",
"model/word_embedding/vocabulary_200dim.pkl",
gpu = 0 , max_doc_len = 500 )
memsum_pubmed = MemSum( "model/memsum-pubmed/model.pt",
"model/word_embedding/vocabulary_200dim.pkl",
gpu = 0 , max_doc_len = 500 )
memsum_gov_report = MemSum( "model/memsum-gov-report/model.pt",
"model/word_embedding/vocabulary_200dim.pkl",
gpu = 0 , max_doc_len = 500 )
```
```python
test_corpus_arxiv = [ json.loads(line) for line in open("data/arxiv/test.jsonl") ]
test_corpus_pubmed = [ json.loads(line) for line in open("data/pubmed/test.jsonl") ]
test_corpus_gov_report = [ json.loads(line) for line in open("data/gov-report/test.jsonl") ]
```
### Evaluation on ROUGE
```python
def evaluate( model, corpus, p_stop, max_extracted_sentences, rouge_cal ):
scores = []
for data in tqdm(corpus):
gold_summary = data["summary"]
extracted_summary = model.extract( [data["text"]], p_stop_thres = p_stop, max_extracted_sentences_per_document = max_extracted_sentences )[0]
score = rouge_cal.score( "\n".join( gold_summary ), "\n".join(extracted_summary) )
scores.append( [score["rouge1"].fmeasure, score["rouge2"].fmeasure, score["rougeLsum"].fmeasure ] )
return np.asarray(scores).mean(axis = 0)
```
```python
evaluate( memsum_arxiv, test_corpus_arxiv, 0.5, 5, rouge_cal )
```
100%|█████████████████████████████████████████████████████████████| 6440/6440 [08:00<00:00, 13.41it/s]
array([0.47946925, 0.19970128, 0.42075852])
```python
evaluate( memsum_pubmed, test_corpus_pubmed, 0.6, 7, rouge_cal )
```
100%|█████████████████████████████████████████████████████████████| 6658/6658 [09:22<00:00, 11.84it/s]
array([0.49260137, 0.22916328, 0.44415123])
```python
evaluate( memsum_gov_report, test_corpus_gov_report, 0.6, 22, rouge_cal )
```
100%|███████████████████████████████████████████████████████████████| 973/973 [04:33<00:00, 3.55it/s]
array([0.59445629, 0.28507926, 0.56677073])
### Summarization Examples
Given a document with a list of sentences, e.g.:
```python
document = test_corpus_pubmed[0]["text"]
```
We can summarize this document extractively by:
```python
extracted_summary = memsum_pubmed.extract( [ document ],
p_stop_thres = 0.6,
max_extracted_sentences_per_document = 7
)[0]
extracted_summary
```
['more specifically , we found that pd patients with anxiety were more impaired on the trail making test part b which assessed attentional set - shifting , on both digit span tests which assessed working memory and attention , and to a lesser extent on the logical memory test which assessed memory and new verbal learning compared to pd patients without anxiety . taken together ,',
'this study is the first to directly compare cognition between pd patients with and without anxiety .',
'results from this study showed selective verbal memory deficits in rpd patients with anxiety compared to rpd without anxiety , whereas lpd patients with anxiety had greater attentional / working memory deficits compared to lpd without anxiety .',
'given that research on healthy young adults suggests that anxiety reduces processing capacity and impairs processing efficiency , especially in the central executive and attentional systems of working memory [ 26 , 27 ] , we hypothesized that pd patients with anxiety would show impairments in attentional set - shifting and working memory compared to pd patients without anxiety .',
'the findings confirmed our hypothesis that anxiety negatively influences attentional set - shifting and working memory in pd .',
'seventeen pd patients with anxiety and thirty - three pd patients without anxiety were included in this study ( see table 1 ) .']
```python
```
We can also get the indices of the extracted sentences in the original document:
```python
extracted_summary_batch, extracted_indices_batch = memsum_pubmed.extract( [ document ],
p_stop_thres = 0.6,
max_extracted_sentences_per_document = 7,
return_sentence_position=1
)
```
```python
extracted_summary_batch[0]
```
['more specifically , we found that pd patients with anxiety were more impaired on the trail making test part b which assessed attentional set - shifting , on both digit span tests which assessed working memory and attention , and to a lesser extent on the logical memory test which assessed memory and new verbal learning compared to pd patients without anxiety . taken together ,',
'this study is the first to directly compare cognition between pd patients with and without anxiety .',
'results from this study showed selective verbal memory deficits in rpd patients with anxiety compared to rpd without anxiety , whereas lpd patients with anxiety had greater attentional / working memory deficits compared to lpd without anxiety .',
'given that research on healthy young adults suggests that anxiety reduces processing capacity and impairs processing efficiency , especially in the central executive and attentional systems of working memory [ 26 , 27 ] , we hypothesized that pd patients with anxiety would show impairments in attentional set - shifting and working memory compared to pd patients without anxiety .',
'the findings confirmed our hypothesis that anxiety negatively influences attentional set - shifting and working memory in pd .',
'seventeen pd patients with anxiety and thirty - three pd patients without anxiety were included in this study ( see table 1 ) .']
```python
extracted_indices_batch[0]
```
[50, 48, 70, 14, 49, 16]
```python
```
## Training MemSum
Please refer to the documentation [Training_Pipeline.md](Training_Pipeline.md) for the complete pipeline of training MemSum on custom dataset.
You can also directly run the training pipeline on google colab: <a href="https://colab.research.google.com/github/nianlonggu/MemSum/blob/main/Training_Pipeline.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>
```python
```
## Updates
### Update 09-02-2023: Released the dataset for human evaluation (comparing MemSum with NeuSum).
Data is available in folder human_eval_results/. It recorded the samples we used for human evaluation and records of participants' labelling.
Released a colab notebook that contained the interface for conducting human evaluation. This can be used for reproducibility test.
Documentation: [MemSum_Human_Evaluation.md](MemSum_Human_Evaluation.md)
Run it on google colab (recommended): <a href="https://colab.research.google.com/github/nianlonggu/MemSum/blob/main/MemSum_Human_Evaluation.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>
### Update 28-07-2022: Code for obtaining the greedy summary of a document
```python
from data_preprocessing.utils import greedy_extract
import json
test_corpus_custom_data = [ json.loads(line) for line in open("data/custom_data/test.jsonl")]
example_data = test_corpus_custom_data[0]
```
```python
example_data.keys()
```
dict_keys(['text', 'summary'])
We can extract the oracle summary by calling the function greedy_extract and set beamsearch_size = 1
```python
greedy_extract( example_data["text"], example_data["summary"], beamsearch_size = 1 )[0]
```
[[50, 13, 41, 24, 31, 0, 3, 48], 0.4563635838327488]
Here the first element is a list of sentence indices in the document, the second element is the avarge of Rouge F1 scores.
```python
```
### References
When using our code or models for your application, please cite the following paper:
```
@inproceedings{gu-etal-2022-memsum,
title = "{M}em{S}um: Extractive Summarization of Long Documents Using Multi-Step Episodic {M}arkov Decision Processes",
author = "Gu, Nianlong and
Ash, Elliott and
Hahnloser, Richard",
booktitle = "Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
month = may,
year = "2022",
address = "Dublin, Ireland",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.acl-long.450",
pages = "6507--6522",
abstract = "We introduce MemSum (Multi-step Episodic Markov decision process extractive SUMmarizer), a reinforcement-learning-based extractive summarizer enriched at each step with information on the current extraction history. When MemSum iteratively selects sentences into the summary, it considers a broad information set that would intuitively also be used by humans in this task: 1) the text content of the sentence, 2) the global text context of the rest of the document, and 3) the extraction history consisting of the set of sentences that have already been extracted. With a lightweight architecture, MemSum obtains state-of-the-art test-set performance (ROUGE) in summarizing long documents taken from PubMed, arXiv, and GovReport. Ablation studies demonstrate the importance of local, global, and history information. A human evaluation confirms the high quality and low redundancy of the generated summaries, stemming from MemSum{'}s awareness of extraction history.",
}
```
| [
"SUMMARIZATION"
] | Non_BioNLP |
SQAI/bge-embedding-model2 | SQAI | sentence-similarity | [
"sentence-transformers",
"safetensors",
"bert",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:1865",
"loss:MatryoshkaLoss",
"loss:MultipleNegativesRankingLoss",
"en",
"arxiv:1908.10084",
"arxiv:2205.13147",
"arxiv:1705.00652",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,719,879,752,000 | 2024-07-02T00:22:56 | 48 | 0 | ---
base_model: SQAI/bge-embedding-model
datasets: []
language:
- en
library_name: sentence-transformers
license: apache-2.0
metrics:
- cosine_accuracy@1
- cosine_accuracy@3
- cosine_accuracy@5
- cosine_accuracy@10
- cosine_precision@1
- cosine_precision@3
- cosine_precision@5
- cosine_precision@10
- cosine_recall@1
- cosine_recall@3
- cosine_recall@5
- cosine_recall@10
- cosine_ndcg@10
- cosine_mrr@10
- cosine_map@100
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:1865
- loss:MatryoshkaLoss
- loss:MultipleNegativesRankingLoss
widget:
- source_sentence: threshold.highLuxThreshold
sentences:
- '"Can you provide the timestamp of the last update to the threshold settings,
and detail any faults in the lux module related to light level sensing and control
for the streetlight on this specific street name? I also want to know the longitude
of the streetlight. And also, can you tell me what type of dimming schedule is
applied to the streetlight, the type of port used for its dimming controls, and
the total energy it has consumed, recorded in kilowatt-hours. Lastly, could you
also provide the timestamp of the recorded streetlighting error, and confirm the
status of the relay responsible for turning this streetlight on and off, as I
am suspecting it might be sticking?"'
- '"Can you provide me with the unique streetlight identifier, upper lux level for
managing light intensity, a brief description, and the delta or height of the
grid area occupied by a group of streetlights? Also, can you note the AC voltage
supply for these streetlights, any issues with communication related to their
lux sensors, and the count of how many times each streetlight has been switched
on? Please ensure that the data is constrained to just those that can be determined
with the unique streetlight identifier I provided."'
- '"What was the last recorded data or action timestamp of the streetlight located
at the specific longitude, and in which time zone is it situated? Could you also
provide information on its default dimming level and the maximum power usage threshold
above which indicates potential faults? Are there any identified faults in the
lux module impacting light level sensing and control? Additionally, what are the
minimum longitude and delta or height for the grid area occupied by this group
of streetlights and could you specify the network time received from the central
control for synchronization purposes?"'
- source_sentence: asset.geoZone
sentences:
- '"Could you check the status of the streetlight with the unique identifier, located
on the named street, specifically looking at any records of complete loss of power
which could indicate supply issues or damage? Also, could you provide details
on the instances where the voltage under load is lower than expected, as well
as instances of lower than expected power consumption, which could signal potential
electrical or hardware issues? I''m also interested in understanding if there
are any faults in our link control mechanism managing multiple streetlights. Additionally,
could you tell me the current drawn by this specific streetlight when it was lower
than expected and the current dimming level of the streetlight in operation? Lastly,
could you specify the maximum safe voltage under load conditions for this light
and verify whether its broadcast subscription used for receiving control signals
is doing fine?"'
- '"Can you provide me with the details regarding a specific streetlight on Main
Street, particularly the minimum current level below which it''s considered abnormal,
its power factor indicating efficient power usage, total operational hours logged,
any incidences where power consumption was higher than expected possibly due to
potential faults, its geoZone, X-coordinate in the grid layout, minimum operational
voltage under load conditions, minimum load current that indicates suboptimal
performance, and the timestamp of the last update made to the threshold settings?"'
- '"What is the width and height of the grid area occupied by the group of streetlights,
type of port used for dimming controls, power consumption levels, and what is
the safety of the current exceeded on the streetlight? Besides, could you explain
the high power factor indicating potential overloads or capacitive imbalances?"'
- source_sentence: errors.deviceId
sentences:
- '"Can you show me a report of all the streetlights with a unique identifier, which
have an internal temperature indicating abnormal operating conditions such as
voltage supplied being below the safe level, and operating temperature below expected
limit possibly due to environmental conditions? Can this report also include instances
of faults in link control mechanism managing multiple streetlights and cases of
open circuit in the relay preventing normal operation?"'
- '"Could you provide information about the streetlight on ''specific street name'',
specifically concerning its current drawn which appears to be lower than expected,
potential issues in the link control mechanism that manages multiple streetlights,
whether its operating temperature exceeds safe limits thus risking damage, and
if its power output is lower than expected? Also, could you let me know at what
interval this streetlight sends data reports and inform about any other issues
detected, particularly when the current is below the expected range?"'
- '"What is the minimum power usage level below which it is considered abnormal
for our ''Main Street Lamps'' group of streetlights, which are described as a
series of LED lamps installed along the main town stretch, and what could be the
reasons if the power consumption is lower than expected, possibly due to hardware
issues? Also, could you give me the description on what means when intermittent
flashing of the streetlight occurs, indicating instability and tell me about the
strength of the wireless signal received by the streetlight''s communication module.
Could you confirm what control mode switch identifier we should use for changing
streetlight settings and the highest power factor that is considered optimal for
streetlight efficiency? Additionally, we discovered issues with group management
of streetlights via our central control system, and we would like to know the
time taken for the streetlight to activate or light up from the command."'
- source_sentence: threshold.lowLoadVoltage
sentences:
- '"Could you please show me the latest data recorded or action performed by the
streetlight, specifically highlighting the control mode switch identifier used
for changing its settings, the type of DALI dimming protocol it uses, and the
type of port used for its dimming controls? Furthermore, has there been any intermittent
flashing indicating instability? Also, could you provide data on its minimum operational
voltage under load conditions, and let me know if its power consumption is lower
than expected due to potential hardware issues?"
'
- '"Can the operator managing the streetlight provide the timestamp of the latest
data recorded or action performed by the streetlight, details on the minimum operational
voltage under load conditions, the current issues with the driver that powers
and controls the streetlight, why the power output is lower than expected for
the streetlight, and what is the maximum latitude of the geographic area covered
by this group of streetlights?"'
- '"Can you provide a report that shows all the streetlights in a grid layout with
Y-coordinate information, indicating whether their control mode setting is on
automated or manual, their minimum current level, and instances of communication
issues between the streetlight''s driver and the control system, as well as instances
when the operating temperature fell below expected limits, possibly due to environmental
conditions?"'
- source_sentence: errors.controllerFault.lowLoadCurrent
sentences:
- '"Can you provide me with the current status of the streetlight on ''street name'',
specifically in relation to its voltage under load, whether it''s lower than expected
and how that might be indicating potential electrical issues? Could you also give
me insight into the current drawn by the streetlight, whether or not the relay
is currently on or off, and if there are any faults in the lux module that may
affect light level sensing and control? Moreover, could you tell me the type of
dimming schedule applied, the ambient light level detected in lux, the total energy
consumed so far recorded in kilowatt-hours, and the lower voltage threshold for
this streetlight''s efficient operation?"'
- '"Can you provide a detailed report for the streetlight on [Name of the street
for the streetlight in error]? The report should include the timestamp of the
last recorded error, synchronization time received from the central control, the
dimming schedule type we''re currently using, and both minimum operational and
maximum safe voltage under load conditions. Also, indicate the time of the last
action was recorded and if there are any reported faults in the metering components
affecting data reporting. Can you also specify the port type used for dimming
controls and whether the power consumption has been unusually low due to potential
hardware issues?"'
- '"Can you show me the current status of the relay in the streetlights located
at the X-coordinate grid, highlighting any faults in the lux module that might
be affecting light level sensing and control? Also, could you provide information
on the current dimming level of these streetlights in operation, the type of dimming
schedule applied, and whether the voltage is within the upper limit considered
safe and efficient for their operation?"'
model-index:
- name: BGE base Financial Matryoshka
results:
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 768
type: dim_768
metrics:
- type: cosine_accuracy@1
value: 0.0
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.0
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.0
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.014423076923076924
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.0
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.0
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.0
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.0014423076923076926
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.0
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.0
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.0
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.014423076923076924
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.004284253930989665
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.001549145299145299
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.005857063109582476
name: Cosine Map@100
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 512
type: dim_512
metrics:
- type: cosine_accuracy@1
value: 0.0
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.0
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.0
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.014423076923076924
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.0
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.0
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.0
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.0014423076923076926
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.0
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.0
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.0
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.014423076923076924
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.004284253930989665
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.001549145299145299
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.005857063109582476
name: Cosine Map@100
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 256
type: dim_256
metrics:
- type: cosine_accuracy@1
value: 0.0
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.0
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.0
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.014423076923076924
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.0
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.0
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.0
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.0014423076923076926
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.0
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.0
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.0
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.014423076923076924
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.0043536523979211435
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.0016159188034188035
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.005708010488423065
name: Cosine Map@100
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 128
type: dim_128
metrics:
- type: cosine_accuracy@1
value: 0.0
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.0
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.0
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.009615384615384616
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.0
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.0
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.0
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.0009615384615384616
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.0
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.0
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.0
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.009615384615384616
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.0030498236971024735
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.001221001221001221
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.005185692544152747
name: Cosine Map@100
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 64
type: dim_64
metrics:
- type: cosine_accuracy@1
value: 0.0
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.0
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.0
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.019230769230769232
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.0
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.0
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.0
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.0019230769230769232
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.0
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.0
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.0
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.019230769230769232
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.005956216500485246
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.0023027319902319903
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.0051874402718147935
name: Cosine Map@100
---
# BGE base Financial Matryoshka
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [SQAI/bge-embedding-model](https://huggingface.co/SQAI/bge-embedding-model). It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [SQAI/bge-embedding-model](https://huggingface.co/SQAI/bge-embedding-model) <!-- at revision 9a9bc3f795ddfc56610a621b37aa077ae0653fa4 -->
- **Maximum Sequence Length:** 512 tokens
- **Output Dimensionality:** 384 tokens
- **Similarity Function:** Cosine Similarity
<!-- - **Training Dataset:** Unknown -->
- **Language:** en
- **License:** apache-2.0
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel
(1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("SQAI/bge-embedding-model2")
# Run inference
sentences = [
'errors.controllerFault.lowLoadCurrent',
'"Can you provide me with the current status of the streetlight on \'street name\', specifically in relation to its voltage under load, whether it\'s lower than expected and how that might be indicating potential electrical issues? Could you also give me insight into the current drawn by the streetlight, whether or not the relay is currently on or off, and if there are any faults in the lux module that may affect light level sensing and control? Moreover, could you tell me the type of dimming schedule applied, the ambient light level detected in lux, the total energy consumed so far recorded in kilowatt-hours, and the lower voltage threshold for this streetlight\'s efficient operation?"',
'"Can you show me the current status of the relay in the streetlights located at the X-coordinate grid, highlighting any faults in the lux module that might be affecting light level sensing and control? Also, could you provide information on the current dimming level of these streetlights in operation, the type of dimming schedule applied, and whether the voltage is within the upper limit considered safe and efficient for their operation?"',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
## Evaluation
### Metrics
#### Information Retrieval
* Dataset: `dim_768`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.0 |
| cosine_accuracy@3 | 0.0 |
| cosine_accuracy@5 | 0.0 |
| cosine_accuracy@10 | 0.0144 |
| cosine_precision@1 | 0.0 |
| cosine_precision@3 | 0.0 |
| cosine_precision@5 | 0.0 |
| cosine_precision@10 | 0.0014 |
| cosine_recall@1 | 0.0 |
| cosine_recall@3 | 0.0 |
| cosine_recall@5 | 0.0 |
| cosine_recall@10 | 0.0144 |
| cosine_ndcg@10 | 0.0043 |
| cosine_mrr@10 | 0.0015 |
| **cosine_map@100** | **0.0059** |
#### Information Retrieval
* Dataset: `dim_512`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.0 |
| cosine_accuracy@3 | 0.0 |
| cosine_accuracy@5 | 0.0 |
| cosine_accuracy@10 | 0.0144 |
| cosine_precision@1 | 0.0 |
| cosine_precision@3 | 0.0 |
| cosine_precision@5 | 0.0 |
| cosine_precision@10 | 0.0014 |
| cosine_recall@1 | 0.0 |
| cosine_recall@3 | 0.0 |
| cosine_recall@5 | 0.0 |
| cosine_recall@10 | 0.0144 |
| cosine_ndcg@10 | 0.0043 |
| cosine_mrr@10 | 0.0015 |
| **cosine_map@100** | **0.0059** |
#### Information Retrieval
* Dataset: `dim_256`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.0 |
| cosine_accuracy@3 | 0.0 |
| cosine_accuracy@5 | 0.0 |
| cosine_accuracy@10 | 0.0144 |
| cosine_precision@1 | 0.0 |
| cosine_precision@3 | 0.0 |
| cosine_precision@5 | 0.0 |
| cosine_precision@10 | 0.0014 |
| cosine_recall@1 | 0.0 |
| cosine_recall@3 | 0.0 |
| cosine_recall@5 | 0.0 |
| cosine_recall@10 | 0.0144 |
| cosine_ndcg@10 | 0.0044 |
| cosine_mrr@10 | 0.0016 |
| **cosine_map@100** | **0.0057** |
#### Information Retrieval
* Dataset: `dim_128`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.0 |
| cosine_accuracy@3 | 0.0 |
| cosine_accuracy@5 | 0.0 |
| cosine_accuracy@10 | 0.0096 |
| cosine_precision@1 | 0.0 |
| cosine_precision@3 | 0.0 |
| cosine_precision@5 | 0.0 |
| cosine_precision@10 | 0.001 |
| cosine_recall@1 | 0.0 |
| cosine_recall@3 | 0.0 |
| cosine_recall@5 | 0.0 |
| cosine_recall@10 | 0.0096 |
| cosine_ndcg@10 | 0.003 |
| cosine_mrr@10 | 0.0012 |
| **cosine_map@100** | **0.0052** |
#### Information Retrieval
* Dataset: `dim_64`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.0 |
| cosine_accuracy@3 | 0.0 |
| cosine_accuracy@5 | 0.0 |
| cosine_accuracy@10 | 0.0192 |
| cosine_precision@1 | 0.0 |
| cosine_precision@3 | 0.0 |
| cosine_precision@5 | 0.0 |
| cosine_precision@10 | 0.0019 |
| cosine_recall@1 | 0.0 |
| cosine_recall@3 | 0.0 |
| cosine_recall@5 | 0.0 |
| cosine_recall@10 | 0.0192 |
| cosine_ndcg@10 | 0.006 |
| cosine_mrr@10 | 0.0023 |
| **cosine_map@100** | **0.0052** |
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### Unnamed Dataset
* Size: 1,865 training samples
* Columns: <code>positive</code> and <code>anchor</code>
* Approximate statistics based on the first 1000 samples:
| | positive | anchor |
|:--------|:---------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
| type | string | string |
| details | <ul><li>min: 5 tokens</li><li>mean: 7.68 tokens</li><li>max: 14 tokens</li></ul> | <ul><li>min: 17 tokens</li><li>mean: 89.79 tokens</li><li>max: 187 tokens</li></ul> |
* Samples:
| positive | anchor |
|:----------------------------------------------|:----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| <code>threshold.lowLoadVoltage</code> | <code>"What is the maximum current level above which it is considered unsafe for a specific streetlight in my area, what is the minimum longitude of the geographic area this streetlight covers, is this streetlight's control mode automated or manually controlled, also, can you provide the delta or width of the grid area occupied by this group of streetlights, what is the level of AC voltage supply to this streetlight, what's the lower voltage threshold below which this streetlight may not operate efficiently, how many times has this streetlight been switched on, what is the minimum operational voltage under load conditions, and finally, what is the latitude of this streetlight?"</code> |
| <code>asset.id</code> | <code>"Could you please tell me the scheduled dimming settings for the string stored streetlights, troubleshoot why these streetlights remain on during daylight hours, and confirm if this could be due to sensor faults? Also, I'd like to know the identifier for the parent group to which this group of streetlights belongs, and the IMEI number of the streetlight device."</code> |
| <code>errors.controllerFault.highPower</code> | <code>"Can you provide an analysis of the efficiency of power usage by examining the power factor of the streetlights, especially in areas of the grid with high Y-coordinates, highlight instances where power consumption is significantly higher than expected which may indicate faults, identify situations where voltage under load is above safe levels, and assess if there are any problems with our central control system's ability to manage streetlight groups?"</code> |
* Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
```json
{
"loss": "MultipleNegativesRankingLoss",
"matryoshka_dims": [
384,
256,
128,
64
],
"matryoshka_weights": [
1,
1,
1,
1
],
"n_dims_per_step": -1
}
```
### Evaluation Dataset
#### Unnamed Dataset
* Size: 208 evaluation samples
* Columns: <code>positive</code> and <code>anchor</code>
* Approximate statistics based on the first 1000 samples:
| | positive | anchor |
|:--------|:---------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|
| type | string | string |
| details | <ul><li>min: 5 tokens</li><li>mean: 7.55 tokens</li><li>max: 14 tokens</li></ul> | <ul><li>min: 19 tokens</li><li>mean: 90.69 tokens</li><li>max: 187 tokens</li></ul> |
* Samples:
| positive | anchor |
|:---------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| <code>log.controlModeSwitch</code> | <code>"Can you provide the control mode switch identifier used for changing the default dimming level set for a specific group of streetlights, identified by their unique identifier, considering the time taken for the streetlight to activate or light up from the command, and possibly troubleshoot why the power consumption is lower than expected which could be due to hardware issues, quite possibly due to the relay responsible for turning the streetlight on and off sticking?"</code> |
| <code>errors.controllerFault.luxModuleFault</code> | <code>"Can you provide the timestamp of the last update to the threshold settings, and detail any faults in the lux module related to light level sensing and control for the streetlight on this specific street name? I also want to know the longitude of the streetlight. And also, can you tell me what type of dimming schedule is applied to the streetlight, the type of port used for its dimming controls, and the total energy it has consumed, recorded in kilowatt-hours. Lastly, could you also provide the timestamp of the recorded streetlighting error, and confirm the status of the relay responsible for turning this streetlight on and off, as I am suspecting it might be sticking?"</code> |
| <code>threshold.lowLoadCurrent</code> | <code>"What is the maximum safe voltage under load conditions for the city's streetlights, and do we possess the necessary rights to link these streetlights for synchronized control? Could you provide me with the timestamp of the latest data or action performed by our streetlights, and tell me the lower lux level threshold at which we would need to consider additional lighting? How often does each streetlight send a data report in normal operation, and what is the minimum load current level where we might start seeing suboptimal functioning? Have we been experiencing any problems with managing groups of streetlights via the central control system? Also, has there been any instances where the current under load was excessively high, indicating possible overloads, or situations where the operation temperature was belo normal limits due to environmental conditions? Lastly, have there been any noted communication issues between the streetlight's driver and the control system?"</code> |
* Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
```json
{
"loss": "MultipleNegativesRankingLoss",
"matryoshka_dims": [
384,
256,
128,
64
],
"matryoshka_weights": [
1,
1,
1,
1
],
"n_dims_per_step": -1
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: epoch
- `per_device_train_batch_size`: 32
- `per_device_eval_batch_size`: 16
- `gradient_accumulation_steps`: 16
- `learning_rate`: 2e-06
- `weight_decay`: 0.03
- `num_train_epochs`: 200
- `lr_scheduler_type`: cosine
- `warmup_ratio`: 0.2
- `bf16`: True
- `tf32`: True
- `load_best_model_at_end`: True
- `optim`: adamw_torch_fused
- `batch_sampler`: no_duplicates
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: epoch
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 32
- `per_device_eval_batch_size`: 16
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 16
- `eval_accumulation_steps`: None
- `learning_rate`: 2e-06
- `weight_decay`: 0.03
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1.0
- `num_train_epochs`: 200
- `max_steps`: -1
- `lr_scheduler_type`: cosine
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.2
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: True
- `fp16`: False
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: True
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: True
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch_fused
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: False
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `dispatch_batches`: None
- `split_batches`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `batch_sampler`: no_duplicates
- `multi_dataset_batch_sampler`: proportional
</details>
### Training Logs
<details><summary>Click to expand</summary>
| Epoch | Step | Training Loss | loss | dim_128_cosine_map@100 | dim_256_cosine_map@100 | dim_512_cosine_map@100 | dim_64_cosine_map@100 | dim_768_cosine_map@100 |
|:----------:|:------:|:-------------:|:----------:|:----------------------:|:----------------------:|:----------------------:|:---------------------:|:----------------------:|
| 0.2712 | 1 | 13.2713 | - | - | - | - | - | - |
| 0.5424 | 2 | 13.2895 | - | - | - | - | - | - |
| 0.8136 | 3 | 9.9139 | - | - | - | - | - | - |
| 1.0847 | 4 | 5.6117 | - | - | - | - | - | - |
| 1.3559 | 5 | 4.7571 | - | - | - | - | - | - |
| 1.6271 | 6 | 5.5215 | - | - | - | - | - | - |
| 1.8983 | 7 | 5.7945 | - | - | - | - | - | - |
| 2.1695 | 8 | 5.7064 | - | - | - | - | - | - |
| 2.4407 | 9 | 5.6794 | - | - | - | - | - | - |
| 2.7119 | 10 | 5.7384 | - | - | - | - | - | - |
| 2.9831 | 11 | 5.6081 | - | - | - | - | - | - |
| 3.2542 | 12 | 5.5278 | - | - | - | - | - | - |
| 3.5254 | 13 | 5.149 | - | - | - | - | - | - |
| 3.7966 | 14 | 5.5904 | 5.6043 | 0.0081 | 0.0072 | 0.0079 | 0.0055 | 0.0079 |
| 1.0169 | 15 | 3.9458 | - | - | - | - | - | - |
| 1.2881 | 16 | 13.3653 | - | - | - | - | - | - |
| 1.5593 | 17 | 13.4413 | - | - | - | - | - | - |
| 1.8305 | 18 | 9.4188 | - | - | - | - | - | - |
| 2.1017 | 19 | 5.717 | - | - | - | - | - | - |
| 2.3729 | 20 | 5.2455 | - | - | - | - | - | - |
| 2.6441 | 21 | 5.2117 | - | - | - | - | - | - |
| 2.9153 | 22 | 5.5217 | - | - | - | - | - | - |
| 3.1864 | 23 | 5.6725 | - | - | - | - | - | - |
| 3.4576 | 24 | 5.786 | - | - | - | - | - | - |
| 3.7288 | 25 | 5.6507 | - | - | - | - | - | - |
| 4.0 | 26 | 5.7215 | - | - | - | - | - | - |
| 4.2712 | 27 | 5.3999 | - | - | - | - | - | - |
| 4.5424 | 28 | 5.4275 | - | - | - | - | - | - |
| 4.8136 | 29 | 5.7143 | 5.5718 | 0.0082 | 0.0071 | 0.0077 | 0.0052 | 0.0077 |
| 2.0339 | 30 | 4.478 | - | - | - | - | - | - |
| 2.3051 | 31 | 13.1821 | - | - | - | - | - | - |
| 2.5763 | 32 | 13.2473 | - | - | - | - | - | - |
| 2.8475 | 33 | 8.8654 | - | - | - | - | - | - |
| 3.1186 | 34 | 5.3181 | - | - | - | - | - | - |
| 3.3898 | 35 | 5.2091 | - | - | - | - | - | - |
| 3.6610 | 36 | 5.6027 | - | - | - | - | - | - |
| 3.9322 | 37 | 5.6839 | - | - | - | - | - | - |
| 4.2034 | 38 | 5.5955 | - | - | - | - | - | - |
| 4.4746 | 39 | 5.5786 | - | - | - | - | - | - |
| 4.7458 | 40 | 5.4509 | - | - | - | - | - | - |
| 5.0169 | 41 | 5.3361 | - | - | - | - | - | - |
| 5.2881 | 42 | 5.1608 | - | - | - | - | - | - |
| 5.5593 | 43 | 5.4896 | - | - | - | - | - | - |
| 5.8305 | 44 | 5.6466 | 5.5241 | 0.0062 | 0.0070 | 0.0076 | 0.0095 | 0.0076 |
| 3.0508 | 45 | 4.5617 | - | - | - | - | - | - |
| 3.3220 | 46 | 13.0665 | - | - | - | - | - | - |
| 3.5932 | 47 | 13.1848 | - | - | - | - | - | - |
| 3.8644 | 48 | 8.4053 | - | - | - | - | - | - |
| 4.1356 | 49 | 5.2706 | - | - | - | - | - | - |
| 4.4068 | 50 | 5.4269 | - | - | - | - | - | - |
| 4.6780 | 51 | 5.3645 | - | - | - | - | - | - |
| 4.9492 | 52 | 5.3587 | - | - | - | - | - | - |
| 5.2203 | 53 | 5.1047 | - | - | - | - | - | - |
| 5.4915 | 54 | 5.743 | - | - | - | - | - | - |
| 5.7627 | 55 | 5.3754 | - | - | - | - | - | - |
| 6.0339 | 56 | 5.3021 | - | - | - | - | - | - |
| 6.3051 | 57 | 5.6983 | - | - | - | - | - | - |
| 6.5763 | 58 | 5.302 | - | - | - | - | - | - |
| 6.8475 | 59 | 5.4545 | 5.4638 | 0.0060 | 0.0070 | 0.0077 | 0.0094 | 0.0077 |
| 4.0678 | 60 | 5.2213 | - | - | - | - | - | - |
| 4.3390 | 61 | 12.9854 | - | - | - | - | - | - |
| 4.6102 | 62 | 13.207 | - | - | - | - | - | - |
| 4.8814 | 63 | 7.7493 | - | - | - | - | - | - |
| 5.1525 | 64 | 5.3787 | - | - | - | - | - | - |
| 5.4237 | 65 | 4.9406 | - | - | - | - | - | - |
| 5.6949 | 66 | 5.3963 | - | - | - | - | - | - |
| 5.9661 | 67 | 5.3429 | - | - | - | - | - | - |
| 6.2373 | 68 | 5.292 | - | - | - | - | - | - |
| 6.5085 | 69 | 5.6738 | - | - | - | - | - | - |
| 6.7797 | 70 | 5.5927 | - | - | - | - | - | - |
| 7.0508 | 71 | 5.5245 | - | - | - | - | - | - |
| 7.3220 | 72 | 4.8334 | - | - | - | - | - | - |
| 7.5932 | 73 | 5.2015 | - | - | - | - | - | - |
| 7.8644 | 74 | 5.5393 | 5.3954 | 0.0060 | 0.0071 | 0.0078 | 0.0094 | 0.0078 |
| 5.0847 | 75 | 5.6168 | - | - | - | - | - | - |
| 5.3559 | 76 | 12.8678 | - | - | - | - | - | - |
| 5.6271 | 77 | 13.2377 | - | - | - | - | - | - |
| 5.8983 | 78 | 7.1882 | - | - | - | - | - | - |
| 6.1695 | 79 | 5.1293 | - | - | - | - | - | - |
| 6.4407 | 80 | 4.9413 | - | - | - | - | - | - |
| 6.7119 | 81 | 5.1763 | - | - | - | - | - | - |
| 6.9831 | 82 | 4.9512 | - | - | - | - | - | - |
| 7.2542 | 83 | 5.2744 | - | - | - | - | - | - |
| 7.5254 | 84 | 5.0573 | - | - | - | - | - | - |
| 7.7966 | 85 | 5.1938 | - | - | - | - | - | - |
| 8.0678 | 86 | 5.1514 | - | - | - | - | - | - |
| 8.3390 | 87 | 4.9808 | - | - | - | - | - | - |
| 8.6102 | 88 | 4.9983 | - | - | - | - | - | - |
| **8.8814** | **89** | **5.3211** | **5.3268** | **0.0062** | **0.0067** | **0.0075** | **0.0095** | **0.0075** |
| 6.1017 | 90 | 6.1513 | - | - | - | - | - | - |
| 6.3729 | 91 | 12.7972 | - | - | - | - | - | - |
| 6.6441 | 92 | 13.0051 | - | - | - | - | - | - |
| 6.9153 | 93 | 6.551 | - | - | - | - | - | - |
| 7.1864 | 94 | 4.6644 | - | - | - | - | - | - |
| 7.4576 | 95 | 4.8619 | - | - | - | - | - | - |
| 7.7288 | 96 | 5.0812 | - | - | - | - | - | - |
| 8.0 | 97 | 4.758 | - | - | - | - | - | - |
| 8.2712 | 98 | 5.1362 | - | - | - | - | - | - |
| 8.5424 | 99 | 5.5405 | - | - | - | - | - | - |
| 8.8136 | 100 | 5.228 | - | - | - | - | - | - |
| 9.0847 | 101 | 5.1084 | - | - | - | - | - | - |
| 9.3559 | 102 | 5.1574 | - | - | - | - | - | - |
| 9.6271 | 103 | 5.3326 | - | - | - | - | - | - |
| 9.8983 | 104 | 5.34 | 5.2658 | 0.0060 | 0.0066 | 0.0076 | 0.0052 | 0.0076 |
| 7.1186 | 105 | 6.5789 | - | - | - | - | - | - |
| 7.3898 | 106 | 12.7557 | - | - | - | - | - | - |
| 7.6610 | 107 | 13.0203 | - | - | - | - | - | - |
| 7.9322 | 108 | 5.7148 | - | - | - | - | - | - |
| 8.2034 | 109 | 4.7945 | - | - | - | - | - | - |
| 8.4746 | 110 | 4.5926 | - | - | - | - | - | - |
| 8.7458 | 111 | 4.6727 | - | - | - | - | - | - |
| 9.0169 | 112 | 5.0886 | - | - | - | - | - | - |
| 9.2881 | 113 | 5.0562 | - | - | - | - | - | - |
| 9.5593 | 114 | 5.2167 | - | - | - | - | - | - |
| 9.8305 | 115 | 5.048 | - | - | - | - | - | - |
| 10.1017 | 116 | 4.7765 | - | - | - | - | - | - |
| 10.3729 | 117 | 4.9875 | - | - | - | - | - | - |
| 10.6441 | 118 | 4.9501 | - | - | - | - | - | - |
| 10.9153 | 119 | 4.756 | 5.2124 | 0.0057 | 0.0065 | 0.0075 | 0.0054 | 0.0075 |
| 8.1356 | 120 | 6.9381 | - | - | - | - | - | - |
| 8.4068 | 121 | 12.7916 | - | - | - | - | - | - |
| 8.6780 | 122 | 12.8517 | - | - | - | - | - | - |
| 8.9492 | 123 | 5.51 | - | - | - | - | - | - |
| 9.2203 | 124 | 4.686 | - | - | - | - | - | - |
| 9.4915 | 125 | 4.6611 | - | - | - | - | - | - |
| 9.7627 | 126 | 5.2767 | - | - | - | - | - | - |
| 10.0339 | 127 | 4.6103 | - | - | - | - | - | - |
| 10.3051 | 128 | 4.957 | - | - | - | - | - | - |
| 10.5763 | 129 | 5.0236 | - | - | - | - | - | - |
| 10.8475 | 130 | 5.0894 | - | - | - | - | - | - |
| 11.1186 | 131 | 4.7025 | - | - | - | - | - | - |
| 11.3898 | 132 | 5.0765 | - | - | - | - | - | - |
| 11.6610 | 133 | 4.6601 | - | - | - | - | - | - |
| 11.9322 | 134 | 4.9064 | 5.1731 | 0.0056 | 0.0060 | 0.0070 | 0.0054 | 0.0070 |
| 9.1525 | 135 | 7.5884 | - | - | - | - | - | - |
| 9.4237 | 136 | 12.679 | - | - | - | - | - | - |
| 9.6949 | 137 | 12.417 | - | - | - | - | - | - |
| 9.9661 | 138 | 5.1632 | - | - | - | - | - | - |
| 10.2373 | 139 | 4.9486 | - | - | - | - | - | - |
| 10.5085 | 140 | 4.6341 | - | - | - | - | - | - |
| 10.7797 | 141 | 4.9664 | - | - | - | - | - | - |
| 11.0508 | 142 | 4.9567 | - | - | - | - | - | - |
| 11.3220 | 143 | 4.7532 | - | - | - | - | - | - |
| 11.5932 | 144 | 5.2556 | - | - | - | - | - | - |
| 11.8644 | 145 | 4.9652 | - | - | - | - | - | - |
| 12.1356 | 146 | 4.8118 | - | - | - | - | - | - |
| 12.4068 | 147 | 4.704 | - | - | - | - | - | - |
| 12.6780 | 148 | 4.8922 | - | - | - | - | - | - |
| 12.9492 | 149 | 4.6571 | 5.1441 | 0.0061 | 0.0055 | 0.0064 | 0.0053 | 0.0064 |
| 10.1695 | 150 | 8.1284 | - | - | - | - | - | - |
| 10.4407 | 151 | 12.5703 | - | - | - | - | - | - |
| 10.7119 | 152 | 11.8696 | - | - | - | - | - | - |
| 10.9831 | 153 | 4.8543 | - | - | - | - | - | - |
| 11.2542 | 154 | 4.8099 | - | - | - | - | - | - |
| 11.5254 | 155 | 4.7009 | - | - | - | - | - | - |
| 11.7966 | 156 | 4.7986 | - | - | - | - | - | - |
| 12.0678 | 157 | 4.7973 | - | - | - | - | - | - |
| 12.3390 | 158 | 4.5529 | - | - | - | - | - | - |
| 12.6102 | 159 | 5.0275 | - | - | - | - | - | - |
| 12.8814 | 160 | 4.6675 | - | - | - | - | - | - |
| 13.1525 | 161 | 4.6538 | - | - | - | - | - | - |
| 13.4237 | 162 | 4.8355 | - | - | - | - | - | - |
| 13.6949 | 163 | 4.6304 | - | - | - | - | - | - |
| 13.9661 | 164 | 4.7047 | 5.1242 | 0.0064 | 0.0054 | 0.0064 | 0.0095 | 0.0064 |
| 11.1864 | 165 | 8.6549 | - | - | - | - | - | - |
| 11.4576 | 166 | 12.4788 | - | - | - | - | - | - |
| 11.7288 | 167 | 11.6425 | - | - | - | - | - | - |
| 12.0 | 168 | 4.5654 | - | - | - | - | - | - |
| 12.2712 | 169 | 4.7016 | - | - | - | - | - | - |
| 12.5424 | 170 | 4.3306 | - | - | - | - | - | - |
| 12.8136 | 171 | 4.9692 | - | - | - | - | - | - |
| 13.0847 | 172 | 4.7557 | - | - | - | - | - | - |
| 13.3559 | 173 | 4.8665 | - | - | - | - | - | - |
| 13.6271 | 174 | 4.8338 | - | - | - | - | - | - |
| 13.8983 | 175 | 4.9221 | - | - | - | - | - | - |
| 14.1695 | 176 | 4.4968 | - | - | - | - | - | - |
| 14.4407 | 177 | 4.6104 | - | - | - | - | - | - |
| 14.7119 | 178 | 4.8449 | - | - | - | - | - | - |
| 14.9831 | 179 | 4.2392 | 5.1123 | 0.0059 | 0.0055 | 0.0065 | 0.0094 | 0.0065 |
| 12.2034 | 180 | 9.4893 | - | - | - | - | - | - |
| 12.4746 | 181 | 12.4241 | - | - | - | - | - | - |
| 12.7458 | 182 | 11.0389 | - | - | - | - | - | - |
| 13.0169 | 183 | 4.7595 | - | - | - | - | - | - |
| 13.2881 | 184 | 4.5408 | - | - | - | - | - | - |
| 13.5593 | 185 | 4.6108 | - | - | - | - | - | - |
| 13.8305 | 186 | 4.5832 | - | - | - | - | - | - |
| 14.1017 | 187 | 4.6741 | - | - | - | - | - | - |
| 14.3729 | 188 | 4.9353 | - | - | - | - | - | - |
| 14.6441 | 189 | 5.0511 | - | - | - | - | - | - |
| 14.9153 | 190 | 4.6575 | - | - | - | - | - | - |
| 15.1864 | 191 | 4.648 | - | - | - | - | - | - |
| 15.4576 | 192 | 4.6224 | - | - | - | - | - | - |
| 15.7288 | 193 | 4.9292 | - | - | - | - | - | - |
| 16.0 | 194 | 3.7805 | 5.1058 | 0.0063 | 0.0057 | 0.0062 | 0.0094 | 0.0062 |
| 13.2203 | 195 | 10.2695 | - | - | - | - | - | - |
| 13.4915 | 196 | 12.5043 | - | - | - | - | - | - |
| 13.7627 | 197 | 10.4795 | - | - | - | - | - | - |
| 14.0339 | 198 | 4.6901 | - | - | - | - | - | - |
| 14.3051 | 199 | 4.6538 | - | - | - | - | - | - |
| 14.5763 | 200 | 4.4736 | - | - | - | - | - | - |
| 14.8475 | 201 | 4.4383 | - | - | - | - | - | - |
| 15.1186 | 202 | 5.0382 | - | - | - | - | - | - |
| 15.3898 | 203 | 4.5636 | - | - | - | - | - | - |
| 15.6610 | 204 | 4.8089 | - | - | - | - | - | - |
| 15.9322 | 205 | 4.4746 | - | - | - | - | - | - |
| 16.2034 | 206 | 4.5876 | - | - | - | - | - | - |
| 16.4746 | 207 | 4.4972 | - | - | - | - | - | - |
| 16.7458 | 208 | 4.8569 | - | - | - | - | - | - |
| 17.0169 | 209 | 3.5883 | 5.1031 | 0.0059 | 0.0057 | 0.0061 | 0.0095 | 0.0061 |
| 14.2373 | 210 | 10.8988 | - | - | - | - | - | - |
| 14.5085 | 211 | 12.4944 | - | - | - | - | - | - |
| 14.7797 | 212 | 10.1041 | - | - | - | - | - | - |
| 15.0508 | 213 | 4.8811 | - | - | - | - | - | - |
| 15.3220 | 214 | 4.6292 | - | - | - | - | - | - |
| 15.5932 | 215 | 4.4828 | - | - | - | - | - | - |
| 15.8644 | 216 | 4.7588 | - | - | - | - | - | - |
| 16.1356 | 217 | 4.26 | - | - | - | - | - | - |
| 16.4068 | 218 | 4.9124 | - | - | - | - | - | - |
| 16.6780 | 219 | 4.8098 | - | - | - | - | - | - |
| 16.9492 | 220 | 4.4439 | - | - | - | - | - | - |
| 17.2203 | 221 | 4.4824 | - | - | - | - | - | - |
| 17.4915 | 222 | 4.7771 | - | - | - | - | - | - |
| 17.7627 | 223 | 4.5966 | - | - | - | - | - | - |
| 18.0339 | 224 | 3.1409 | 5.1009 | 0.0055 | 0.0057 | 0.0062 | 0.0052 | 0.0062 |
| 15.2542 | 225 | 11.657 | - | - | - | - | - | - |
| 15.5254 | 226 | 12.5032 | - | - | - | - | - | - |
| 15.7966 | 227 | 9.4495 | - | - | - | - | - | - |
| 16.0678 | 228 | 4.7099 | - | - | - | - | - | - |
| 16.3390 | 229 | 4.6049 | - | - | - | - | - | - |
| 16.6102 | 230 | 4.6311 | - | - | - | - | - | - |
| 16.8814 | 231 | 4.7562 | - | - | - | - | - | - |
| 17.1525 | 232 | 4.7195 | - | - | - | - | - | - |
| 17.4237 | 233 | 4.8557 | - | - | - | - | - | - |
| 17.6949 | 234 | 4.8423 | - | - | - | - | - | - |
| 17.9661 | 235 | 4.5764 | - | - | - | - | - | - |
| 18.2373 | 236 | 4.5081 | - | - | - | - | - | - |
| 18.5085 | 237 | 4.7974 | - | - | - | - | - | - |
| 18.7797 | 238 | 4.871 | - | - | - | - | - | - |
| 19.0508 | 239 | 2.8558 | 5.1020 | 0.0054 | 0.0057 | 0.0061 | 0.0054 | 0.0061 |
| 16.2712 | 240 | 12.4297 | - | - | - | - | - | - |
| 16.5424 | 241 | 12.5186 | - | - | - | - | - | - |
| 16.8136 | 242 | 8.8827 | - | - | - | - | - | - |
| 17.0847 | 243 | 4.8406 | - | - | - | - | - | - |
| 17.3559 | 244 | 4.4367 | - | - | - | - | - | - |
| 17.6271 | 245 | 4.5996 | - | - | - | - | - | - |
| 17.8983 | 246 | 4.6692 | - | - | - | - | - | - |
| 18.1695 | 247 | 4.6498 | - | - | - | - | - | - |
| 18.4407 | 248 | 4.7211 | - | - | - | - | - | - |
| 18.7119 | 249 | 4.7675 | - | - | - | - | - | - |
| 18.9831 | 250 | 4.4405 | - | - | - | - | - | - |
| 19.2542 | 251 | 4.6778 | - | - | - | - | - | - |
| 19.5254 | 252 | 4.6674 | - | - | - | - | - | - |
| 19.7966 | 253 | 4.735 | 5.1036 | 0.0054 | 0.0056 | 0.0060 | 0.0054 | 0.0060 |
| 17.0169 | 254 | 3.6188 | - | - | - | - | - | - |
| 17.2881 | 255 | 12.4112 | - | - | - | - | - | - |
| 17.5593 | 256 | 12.5261 | - | - | - | - | - | - |
| 17.8305 | 257 | 8.3408 | - | - | - | - | - | - |
| 18.1017 | 258 | 4.6496 | - | - | - | - | - | - |
| 18.3729 | 259 | 4.7177 | - | - | - | - | - | - |
| 18.6441 | 260 | 4.7956 | - | - | - | - | - | - |
| 18.9153 | 261 | 4.7228 | - | - | - | - | - | - |
| 19.1864 | 262 | 4.6083 | - | - | - | - | - | - |
| 19.4576 | 263 | 4.7985 | - | - | - | - | - | - |
| 19.7288 | 264 | 4.6675 | - | - | - | - | - | - |
| 20.0 | 265 | 4.6353 | - | - | - | - | - | - |
| 20.2712 | 266 | 4.5717 | - | - | - | - | - | - |
| 20.5424 | 267 | 4.4358 | - | - | - | - | - | - |
| 20.8136 | 268 | 4.8288 | 5.1030 | 0.0056 | 0.0057 | 0.0062 | 0.0053 | 0.0062 |
| 18.0339 | 269 | 3.7877 | - | - | - | - | - | - |
| 18.3051 | 270 | 12.4042 | - | - | - | - | - | - |
| 18.5763 | 271 | 12.4793 | - | - | - | - | - | - |
| 18.8475 | 272 | 7.9475 | - | - | - | - | - | - |
| 19.1186 | 273 | 4.5502 | - | - | - | - | - | - |
| 19.3898 | 274 | 4.5565 | - | - | - | - | - | - |
| 19.6610 | 275 | 4.4172 | - | - | - | - | - | - |
| 19.9322 | 276 | 4.5319 | - | - | - | - | - | - |
| 20.2034 | 277 | 4.5635 | - | - | - | - | - | - |
| 20.4746 | 278 | 4.5233 | - | - | - | - | - | - |
| 20.7458 | 279 | 4.6766 | - | - | - | - | - | - |
| 21.0169 | 280 | 4.5863 | - | - | - | - | - | - |
| 21.2881 | 281 | 4.5784 | - | - | - | - | - | - |
| 21.5593 | 282 | 4.7198 | - | - | - | - | - | - |
| 21.8305 | 283 | 4.7383 | 5.1065 | 0.0054 | 0.0056 | 0.0061 | 0.0050 | 0.0061 |
| 19.0508 | 284 | 4.4257 | - | - | - | - | - | - |
| 19.3220 | 285 | 12.3475 | - | - | - | - | - | - |
| 19.5932 | 286 | 12.5168 | - | - | - | - | - | - |
| 19.8644 | 287 | 7.3671 | - | - | - | - | - | - |
| 20.1356 | 288 | 4.3771 | - | - | - | - | - | - |
| 20.4068 | 289 | 4.542 | - | - | - | - | - | - |
| 20.6780 | 290 | 4.3629 | - | - | - | - | - | - |
| 20.9492 | 291 | 4.5474 | - | - | - | - | - | - |
| 21.2203 | 292 | 4.7436 | - | - | - | - | - | - |
| 21.4915 | 293 | 4.5915 | - | - | - | - | - | - |
| 21.7627 | 294 | 4.5522 | - | - | - | - | - | - |
| 22.0339 | 295 | 4.6591 | - | - | - | - | - | - |
| 22.3051 | 296 | 4.6179 | - | - | - | - | - | - |
| 22.5763 | 297 | 4.6086 | - | - | - | - | - | - |
| 22.8475 | 298 | 4.8808 | 5.1083 | 0.0054 | 0.0057 | 0.0062 | 0.0055 | 0.0062 |
| 20.0678 | 299 | 4.7358 | - | - | - | - | - | - |
| 20.3390 | 300 | 12.3209 | - | - | - | - | - | - |
| 20.6102 | 301 | 12.6406 | - | - | - | - | - | - |
| 20.8814 | 302 | 6.7971 | - | - | - | - | - | - |
| 21.1525 | 303 | 4.3723 | - | - | - | - | - | - |
| 21.4237 | 304 | 4.61 | - | - | - | - | - | - |
| 21.6949 | 305 | 4.4624 | - | - | - | - | - | - |
| 21.9661 | 306 | 4.6145 | - | - | - | - | - | - |
| 22.2373 | 307 | 4.5794 | - | - | - | - | - | - |
| 22.5085 | 308 | 4.6625 | - | - | - | - | - | - |
| 22.7797 | 309 | 4.5499 | - | - | - | - | - | - |
| 23.0508 | 310 | 4.5657 | - | - | - | - | - | - |
| 23.3220 | 311 | 4.5896 | - | - | - | - | - | - |
| 23.5932 | 312 | 4.5692 | - | - | - | - | - | - |
| 23.8644 | 313 | 4.93 | 5.1119 | 0.0055 | 0.0057 | 0.0061 | 0.0056 | 0.0061 |
| 21.0847 | 314 | 5.3829 | - | - | - | - | - | - |
| 21.3559 | 315 | 12.3422 | - | - | - | - | - | - |
| 21.6271 | 316 | 12.601 | - | - | - | - | - | - |
| 21.8983 | 317 | 6.5062 | - | - | - | - | - | - |
| 22.1695 | 318 | 4.4681 | - | - | - | - | - | - |
| 22.4407 | 319 | 4.4244 | - | - | - | - | - | - |
| 22.7119 | 320 | 4.4514 | - | - | - | - | - | - |
| 22.9831 | 321 | 4.5469 | - | - | - | - | - | - |
| 23.2542 | 322 | 4.6924 | - | - | - | - | - | - |
| 23.5254 | 323 | 4.682 | - | - | - | - | - | - |
| 23.7966 | 324 | 4.6403 | - | - | - | - | - | - |
| 24.0678 | 325 | 4.6272 | - | - | - | - | - | - |
| 24.3390 | 326 | 4.3605 | - | - | - | - | - | - |
| 24.6102 | 327 | 4.5992 | - | - | - | - | - | - |
| 24.8814 | 328 | 4.6776 | 5.1126 | 0.0053 | 0.0057 | 0.0061 | 0.0056 | 0.0061 |
| 22.1017 | 329 | 5.8504 | - | - | - | - | - | - |
| 22.3729 | 330 | 12.335 | - | - | - | - | - | - |
| 22.6441 | 331 | 12.5779 | - | - | - | - | - | - |
| 22.9153 | 332 | 5.7261 | - | - | - | - | - | - |
| 23.1864 | 333 | 4.5411 | - | - | - | - | - | - |
| 23.4576 | 334 | 4.4783 | - | - | - | - | - | - |
| 23.7288 | 335 | 4.5589 | - | - | - | - | - | - |
| 24.0 | 336 | 4.6305 | - | - | - | - | - | - |
| 24.2712 | 337 | 4.674 | - | - | - | - | - | - |
| 24.5424 | 338 | 4.7455 | - | - | - | - | - | - |
| 24.8136 | 339 | 4.6011 | - | - | - | - | - | - |
| 25.0847 | 340 | 4.5899 | - | - | - | - | - | - |
| 25.3559 | 341 | 4.3981 | - | - | - | - | - | - |
| 25.6271 | 342 | 4.7031 | - | - | - | - | - | - |
| 25.8983 | 343 | 4.68 | 5.1182 | 0.0054 | 0.0057 | 0.0059 | 0.0056 | 0.0059 |
| 23.1186 | 344 | 6.3521 | - | - | - | - | - | - |
| 23.3898 | 345 | 12.2283 | - | - | - | - | - | - |
| 23.6610 | 346 | 12.533 | - | - | - | - | - | - |
| 23.9322 | 347 | 5.2654 | - | - | - | - | - | - |
| 24.2034 | 348 | 4.3667 | - | - | - | - | - | - |
| 24.4746 | 349 | 4.4718 | - | - | - | - | - | - |
| 24.7458 | 350 | 4.6212 | - | - | - | - | - | - |
| 25.0169 | 351 | 4.447 | - | - | - | - | - | - |
| 25.2881 | 352 | 4.6247 | - | - | - | - | - | - |
| 25.5593 | 353 | 5.0093 | - | - | - | - | - | - |
| 25.8305 | 354 | 4.6316 | - | - | - | - | - | - |
| 26.1017 | 355 | 4.6655 | - | - | - | - | - | - |
| 26.3729 | 356 | 4.5964 | - | - | - | - | - | - |
| 26.6441 | 357 | 4.682 | - | - | - | - | - | - |
| 26.9153 | 358 | 4.6375 | 5.1205 | 0.0051 | 0.0056 | 0.0059 | 0.0055 | 0.0059 |
| 24.1356 | 359 | 6.727 | - | - | - | - | - | - |
| 24.4068 | 360 | 12.3706 | - | - | - | - | - | - |
| 24.6780 | 361 | 12.4755 | - | - | - | - | - | - |
| 24.9492 | 362 | 4.623 | - | - | - | - | - | - |
| 25.2203 | 363 | 4.2947 | - | - | - | - | - | - |
| 25.4915 | 364 | 4.3993 | - | - | - | - | - | - |
| 25.7627 | 365 | 4.4148 | - | - | - | - | - | - |
| 26.0339 | 366 | 4.2376 | - | - | - | - | - | - |
| 26.3051 | 367 | 4.6334 | - | - | - | - | - | - |
| 26.5763 | 368 | 4.7007 | - | - | - | - | - | - |
| 26.8475 | 369 | 4.3542 | - | - | - | - | - | - |
| 27.1186 | 370 | 4.7036 | - | - | - | - | - | - |
| 27.3898 | 371 | 4.2382 | - | - | - | - | - | - |
| 27.6610 | 372 | 4.5011 | - | - | - | - | - | - |
| 27.9322 | 373 | 4.6292 | 5.1241 | 0.0051 | 0.0056 | 0.0059 | 0.0056 | 0.0059 |
| 25.1525 | 374 | 7.3562 | - | - | - | - | - | - |
| 25.4237 | 375 | 12.2926 | - | - | - | - | - | - |
| 25.6949 | 376 | 12.1694 | - | - | - | - | - | - |
| 25.9661 | 377 | 4.7183 | - | - | - | - | - | - |
| 26.2373 | 378 | 4.4099 | - | - | - | - | - | - |
| 26.5085 | 379 | 4.3366 | - | - | - | - | - | - |
| 26.7797 | 380 | 4.4848 | - | - | - | - | - | - |
| 27.0508 | 381 | 4.6947 | - | - | - | - | - | - |
| 27.3220 | 382 | 4.5683 | - | - | - | - | - | - |
| 27.5932 | 383 | 4.7691 | - | - | - | - | - | - |
| 27.8644 | 384 | 4.3879 | - | - | - | - | - | - |
| 28.1356 | 385 | 4.3461 | - | - | - | - | - | - |
| 28.4068 | 386 | 4.4756 | - | - | - | - | - | - |
| 28.6780 | 387 | 4.5355 | - | - | - | - | - | - |
| 28.9492 | 388 | 4.4837 | 5.1278 | 0.0052 | 0.0056 | 0.0059 | 0.0054 | 0.0059 |
| 26.1695 | 389 | 7.9407 | - | - | - | - | - | - |
| 26.4407 | 390 | 12.3054 | - | - | - | - | - | - |
| 26.7119 | 391 | 11.6158 | - | - | - | - | - | - |
| 26.9831 | 392 | 4.5724 | - | - | - | - | - | - |
| 27.2542 | 393 | 4.467 | - | - | - | - | - | - |
| 27.5254 | 394 | 4.4395 | - | - | - | - | - | - |
| 27.7966 | 395 | 4.4111 | - | - | - | - | - | - |
| 28.0678 | 396 | 4.5565 | - | - | - | - | - | - |
| 28.3390 | 397 | 4.6063 | - | - | - | - | - | - |
| 28.6102 | 398 | 4.5312 | - | - | - | - | - | - |
| 28.8814 | 399 | 4.5436 | - | - | - | - | - | - |
| 29.1525 | 400 | 4.5366 | - | - | - | - | - | - |
| 29.4237 | 401 | 4.4488 | - | - | - | - | - | - |
| 29.6949 | 402 | 4.5641 | - | - | - | - | - | - |
| 29.9661 | 403 | 4.2491 | 5.1303 | 0.0053 | 0.0057 | 0.0060 | 0.0055 | 0.0060 |
| 27.1864 | 404 | 8.574 | - | - | - | - | - | - |
| 27.4576 | 405 | 12.2836 | - | - | - | - | - | - |
| 27.7288 | 406 | 11.1935 | - | - | - | - | - | - |
| 28.0 | 407 | 4.5464 | - | - | - | - | - | - |
| 28.2712 | 408 | 4.3132 | - | - | - | - | - | - |
| 28.5424 | 409 | 4.3553 | - | - | - | - | - | - |
| 28.8136 | 410 | 4.4679 | - | - | - | - | - | - |
| 29.0847 | 411 | 4.7705 | - | - | - | - | - | - |
| 29.3559 | 412 | 4.5667 | - | - | - | - | - | - |
| 29.6271 | 413 | 4.6547 | - | - | - | - | - | - |
| 29.8983 | 414 | 4.6709 | - | - | - | - | - | - |
| 30.1695 | 415 | 4.784 | - | - | - | - | - | - |
| 30.4407 | 416 | 4.4368 | - | - | - | - | - | - |
| 30.7119 | 417 | 4.6159 | - | - | - | - | - | - |
| 30.9831 | 418 | 4.0117 | 5.1322 | 0.0050 | 0.0057 | 0.0059 | 0.0054 | 0.0059 |
| 28.2034 | 419 | 9.2905 | - | - | - | - | - | - |
| 28.4746 | 420 | 12.2439 | - | - | - | - | - | - |
| 28.7458 | 421 | 10.722 | - | - | - | - | - | - |
| 29.0169 | 422 | 4.6608 | - | - | - | - | - | - |
| 29.2881 | 423 | 4.5196 | - | - | - | - | - | - |
| 29.5593 | 424 | 4.4313 | - | - | - | - | - | - |
| 29.8305 | 425 | 4.513 | - | - | - | - | - | - |
| 30.1017 | 426 | 4.5812 | - | - | - | - | - | - |
| 30.3729 | 427 | 4.5275 | - | - | - | - | - | - |
| 30.6441 | 428 | 4.8022 | - | - | - | - | - | - |
| 30.9153 | 429 | 4.5171 | - | - | - | - | - | - |
| 31.1864 | 430 | 4.5968 | - | - | - | - | - | - |
| 31.4576 | 431 | 4.2145 | - | - | - | - | - | - |
| 31.7288 | 432 | 4.7041 | - | - | - | - | - | - |
| 32.0 | 433 | 3.6187 | 5.1356 | 0.0051 | 0.0057 | 0.0059 | 0.0055 | 0.0059 |
| 29.2203 | 434 | 10.0897 | - | - | - | - | - | - |
| 29.4915 | 435 | 12.2909 | - | - | - | - | - | - |
| 29.7627 | 436 | 10.1362 | - | - | - | - | - | - |
| 30.0339 | 437 | 4.5172 | - | - | - | - | - | - |
| 30.3051 | 438 | 4.3273 | - | - | - | - | - | - |
| 30.5763 | 439 | 4.5272 | - | - | - | - | - | - |
| 30.8475 | 440 | 4.376 | - | - | - | - | - | - |
| 31.1186 | 441 | 4.5803 | - | - | - | - | - | - |
| 31.3898 | 442 | 4.5654 | - | - | - | - | - | - |
| 31.6610 | 443 | 4.5024 | - | - | - | - | - | - |
| 31.9322 | 444 | 4.5889 | - | - | - | - | - | - |
| 32.2034 | 445 | 4.6489 | - | - | - | - | - | - |
| 32.4746 | 446 | 4.4505 | - | - | - | - | - | - |
| 32.7458 | 447 | 4.7026 | - | - | - | - | - | - |
| 33.0169 | 448 | 3.4719 | 5.1368 | 0.0050 | 0.0056 | 0.0059 | 0.0052 | 0.0059 |
| 30.2373 | 449 | 10.7633 | - | - | - | - | - | - |
| 30.5085 | 450 | 12.3203 | - | - | - | - | - | - |
| 30.7797 | 451 | 9.7535 | - | - | - | - | - | - |
| 31.0508 | 452 | 4.7462 | - | - | - | - | - | - |
| 31.3220 | 453 | 4.4271 | - | - | - | - | - | - |
| 31.5932 | 454 | 4.4347 | - | - | - | - | - | - |
| 31.8644 | 455 | 4.6443 | - | - | - | - | - | - |
| 32.1356 | 456 | 4.6344 | - | - | - | - | - | - |
| 32.4068 | 457 | 4.6518 | - | - | - | - | - | - |
| 32.6780 | 458 | 4.6437 | - | - | - | - | - | - |
| 32.9492 | 459 | 4.6168 | - | - | - | - | - | - |
| 33.2203 | 460 | 4.4948 | - | - | - | - | - | - |
| 33.4915 | 461 | 4.5268 | - | - | - | - | - | - |
| 33.7627 | 462 | 4.4844 | - | - | - | - | - | - |
| 34.0339 | 463 | 3.276 | 5.1384 | 0.0051 | 0.0057 | 0.0060 | 0.0053 | 0.0060 |
| 31.2542 | 464 | 11.5311 | - | - | - | - | - | - |
| 31.5254 | 465 | 12.3812 | - | - | - | - | - | - |
| 31.7966 | 466 | 9.1499 | - | - | - | - | - | - |
| 32.0678 | 467 | 4.7032 | - | - | - | - | - | - |
| 32.3390 | 468 | 4.2429 | - | - | - | - | - | - |
| 32.6102 | 469 | 4.549 | - | - | - | - | - | - |
| 32.8814 | 470 | 4.7083 | - | - | - | - | - | - |
| 33.1525 | 471 | 4.5348 | - | - | - | - | - | - |
| 33.4237 | 472 | 4.472 | - | - | - | - | - | - |
| 33.6949 | 473 | 4.5818 | - | - | - | - | - | - |
| 33.9661 | 474 | 4.5534 | - | - | - | - | - | - |
| 34.2373 | 475 | 4.5743 | - | - | - | - | - | - |
| 34.5085 | 476 | 4.54 | - | - | - | - | - | - |
| 34.7797 | 477 | 4.681 | - | - | - | - | - | - |
| 35.0508 | 478 | 2.9902 | 5.1397 | 0.0052 | 0.0057 | 0.0059 | 0.0053 | 0.0059 |
| 32.2712 | 479 | 12.3174 | - | - | - | - | - | - |
| 32.5424 | 480 | 12.2996 | - | - | - | - | - | - |
| 32.8136 | 481 | 8.7153 | - | - | - | - | - | - |
| 33.0847 | 482 | 4.5692 | - | - | - | - | - | - |
| 33.3559 | 483 | 4.3255 | - | - | - | - | - | - |
| 33.6271 | 484 | 4.4515 | - | - | - | - | - | - |
| 33.8983 | 485 | 4.6708 | - | - | - | - | - | - |
| 34.1695 | 486 | 4.2648 | - | - | - | - | - | - |
| 34.4407 | 487 | 4.6268 | - | - | - | - | - | - |
| 34.7119 | 488 | 4.703 | - | - | - | - | - | - |
| 34.9831 | 489 | 4.6269 | - | - | - | - | - | - |
| 35.2542 | 490 | 4.6464 | - | - | - | - | - | - |
| 35.5254 | 491 | 4.4952 | - | - | - | - | - | - |
| 35.7966 | 492 | 4.6097 | 5.1406 | 0.0052 | 0.0058 | 0.0058 | 0.0054 | 0.0058 |
| 33.0169 | 493 | 3.2718 | - | - | - | - | - | - |
| 33.2881 | 494 | 12.3329 | - | - | - | - | - | - |
| 33.5593 | 495 | 12.3503 | - | - | - | - | - | - |
| 33.8305 | 496 | 8.1544 | - | - | - | - | - | - |
| 34.1017 | 497 | 4.4684 | - | - | - | - | - | - |
| 34.3729 | 498 | 4.4062 | - | - | - | - | - | - |
| 34.6441 | 499 | 4.2644 | - | - | - | - | - | - |
| 34.9153 | 500 | 4.5294 | - | - | - | - | - | - |
| 35.1864 | 501 | 4.673 | - | - | - | - | - | - |
| 35.4576 | 502 | 4.4884 | - | - | - | - | - | - |
| 35.7288 | 503 | 4.5989 | - | - | - | - | - | - |
| 36.0 | 504 | 4.6182 | - | - | - | - | - | - |
| 36.2712 | 505 | 4.6487 | - | - | - | - | - | - |
| 36.5424 | 506 | 4.6436 | - | - | - | - | - | - |
| 36.8136 | 507 | 4.6059 | 5.1417 | 0.0051 | 0.0057 | 0.0059 | 0.0052 | 0.0059 |
| 34.0339 | 508 | 3.7589 | - | - | - | - | - | - |
| 34.3051 | 509 | 12.2815 | - | - | - | - | - | - |
| 34.5763 | 510 | 12.5481 | - | - | - | - | - | - |
| 34.8475 | 511 | 7.6339 | - | - | - | - | - | - |
| 35.1186 | 512 | 4.5528 | - | - | - | - | - | - |
| 35.3898 | 513 | 4.3266 | - | - | - | - | - | - |
| 35.6610 | 514 | 4.3093 | - | - | - | - | - | - |
| 35.9322 | 515 | 4.7401 | - | - | - | - | - | - |
| 36.2034 | 516 | 4.523 | - | - | - | - | - | - |
| 36.4746 | 517 | 4.5255 | - | - | - | - | - | - |
| 36.7458 | 518 | 4.5058 | - | - | - | - | - | - |
| 37.0169 | 519 | 4.5614 | - | - | - | - | - | - |
| 37.2881 | 520 | 4.5323 | - | - | - | - | - | - |
| 37.5593 | 521 | 4.5739 | - | - | - | - | - | - |
| 37.8305 | 522 | 4.6501 | 5.1427 | 0.0052 | 0.0058 | 0.0059 | 0.0053 | 0.0059 |
| 35.0508 | 523 | 4.2083 | - | - | - | - | - | - |
| 35.3220 | 524 | 12.2888 | - | - | - | - | - | - |
| 35.5932 | 525 | 12.4709 | - | - | - | - | - | - |
| 35.8644 | 526 | 7.3926 | - | - | - | - | - | - |
| 36.1356 | 527 | 4.4719 | - | - | - | - | - | - |
| 36.4068 | 528 | 4.5033 | - | - | - | - | - | - |
| 36.6780 | 529 | 4.388 | - | - | - | - | - | - |
| 36.9492 | 530 | 4.5606 | - | - | - | - | - | - |
| 37.2203 | 531 | 4.6936 | - | - | - | - | - | - |
| 37.4915 | 532 | 4.6008 | - | - | - | - | - | - |
| 37.7627 | 533 | 4.6973 | - | - | - | - | - | - |
| 38.0339 | 534 | 4.4194 | - | - | - | - | - | - |
| 38.3051 | 535 | 4.5616 | - | - | - | - | - | - |
| 38.5763 | 536 | 4.6307 | - | - | - | - | - | - |
| 38.8475 | 537 | 4.8322 | 5.1442 | 0.0051 | 0.0057 | 0.0059 | 0.0053 | 0.0059 |
| 36.0678 | 538 | 4.8388 | - | - | - | - | - | - |
| 36.3390 | 539 | 12.2334 | - | - | - | - | - | - |
| 36.6102 | 540 | 12.4205 | - | - | - | - | - | - |
| 36.8814 | 541 | 6.9051 | - | - | - | - | - | - |
| 37.1525 | 542 | 4.6011 | - | - | - | - | - | - |
| 37.4237 | 543 | 4.4701 | - | - | - | - | - | - |
| 37.6949 | 544 | 4.421 | - | - | - | - | - | - |
| 37.9661 | 545 | 4.6877 | - | - | - | - | - | - |
| 38.2373 | 546 | 4.6348 | - | - | - | - | - | - |
| 38.5085 | 547 | 4.5822 | - | - | - | - | - | - |
| 38.7797 | 548 | 4.5697 | - | - | - | - | - | - |
| 39.0508 | 549 | 4.3118 | - | - | - | - | - | - |
| 39.3220 | 550 | 4.5131 | - | - | - | - | - | - |
| 39.5932 | 551 | 4.4879 | - | - | - | - | - | - |
| 39.8644 | 552 | 4.5945 | 5.1429 | 0.0052 | 0.0056 | 0.0059 | 0.0054 | 0.0059 |
| 37.0847 | 553 | 5.4083 | - | - | - | - | - | - |
| 37.3559 | 554 | 12.2092 | - | - | - | - | - | - |
| 37.6271 | 555 | 12.5043 | - | - | - | - | - | - |
| 37.8983 | 556 | 6.1239 | - | - | - | - | - | - |
| 38.1695 | 557 | 4.2932 | - | - | - | - | - | - |
| 38.4407 | 558 | 4.3845 | - | - | - | - | - | - |
| 38.7119 | 559 | 4.5619 | - | - | - | - | - | - |
| 38.9831 | 560 | 4.6936 | - | - | - | - | - | - |
| 39.2542 | 561 | 4.6636 | - | - | - | - | - | - |
| 39.5254 | 562 | 4.7964 | - | - | - | - | - | - |
| 39.7966 | 563 | 4.613 | - | - | - | - | - | - |
| 40.0678 | 564 | 4.5856 | - | - | - | - | - | - |
| 40.3390 | 565 | 4.4605 | - | - | - | - | - | - |
| 40.6102 | 566 | 4.5461 | - | - | - | - | - | - |
| 40.8814 | 567 | 4.7145 | 5.1454 | 0.0052 | 0.0056 | 0.0059 | 0.0052 | 0.0059 |
| 38.1017 | 568 | 5.8311 | - | - | - | - | - | - |
| 38.3729 | 569 | 12.2142 | - | - | - | - | - | - |
| 38.6441 | 570 | 12.4489 | - | - | - | - | - | - |
| 38.9153 | 571 | 5.7328 | - | - | - | - | - | - |
| 39.1864 | 572 | 4.4402 | - | - | - | - | - | - |
| 39.4576 | 573 | 4.1806 | - | - | - | - | - | - |
| 39.7288 | 574 | 4.6327 | - | - | - | - | - | - |
| 40.0 | 575 | 4.2768 | - | - | - | - | - | - |
| 40.2712 | 576 | 4.4669 | - | - | - | - | - | - |
| 40.5424 | 577 | 4.8094 | - | - | - | - | - | - |
| 40.8136 | 578 | 4.5773 | - | - | - | - | - | - |
| 41.0847 | 579 | 4.439 | - | - | - | - | - | - |
| 41.3559 | 580 | 4.5718 | - | - | - | - | - | - |
| 41.6271 | 581 | 4.5955 | - | - | - | - | - | - |
| 41.8983 | 582 | 4.5043 | 5.1443 | 0.0051 | 0.0056 | 0.0059 | 0.0054 | 0.0059 |
| 39.1186 | 583 | 6.359 | - | - | - | - | - | - |
| 39.3898 | 584 | 12.212 | - | - | - | - | - | - |
| 39.6610 | 585 | 12.538 | - | - | - | - | - | - |
| 39.9322 | 586 | 5.0971 | - | - | - | - | - | - |
| 40.2034 | 587 | 4.4783 | - | - | - | - | - | - |
| 40.4746 | 588 | 4.394 | - | - | - | - | - | - |
| 40.7458 | 589 | 4.4847 | - | - | - | - | - | - |
| 41.0169 | 590 | 4.4116 | - | - | - | - | - | - |
| 41.2881 | 591 | 4.3979 | - | - | - | - | - | - |
| 41.5593 | 592 | 4.6652 | - | - | - | - | - | - |
| 41.8305 | 593 | 4.3939 | - | - | - | - | - | - |
| 42.1017 | 594 | 4.5555 | - | - | - | - | - | - |
| 42.3729 | 595 | 4.4966 | - | - | - | - | - | - |
| 42.6441 | 596 | 4.6267 | - | - | - | - | - | - |
| 42.9153 | 597 | 4.5834 | 5.1446 | 0.0051 | 0.0057 | 0.0058 | 0.0052 | 0.0058 |
| 40.1356 | 598 | 6.7009 | - | - | - | - | - | - |
| 40.4068 | 599 | 12.2755 | - | - | - | - | - | - |
| 40.6780 | 600 | 12.4465 | 5.1447 | 0.0052 | 0.0057 | 0.0059 | 0.0052 | 0.0059 |
* The bold row denotes the saved checkpoint.
</details>
### Framework Versions
- Python: 3.10.12
- Sentence Transformers: 3.0.1
- Transformers: 4.41.2
- PyTorch: 2.1.2+cu121
- Accelerate: 0.31.0
- Datasets: 2.19.1
- Tokenizers: 0.19.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### MatryoshkaLoss
```bibtex
@misc{kusupati2024matryoshka,
title={Matryoshka Representation Learning},
author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
year={2024},
eprint={2205.13147},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
#### MultipleNegativesRankingLoss
```bibtex
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
MBZUAI/bactrian-x-bloom-7b1-lora | MBZUAI | null | [
"arxiv:2305.15011",
"license:mit",
"region:us"
] | 1,683,744,074,000 | 2023-06-11T10:11:59 | 0 | 0 | ---
license: mit
---
#### Current Training Steps: 100,000
This repo contains a low-rank adapter (LoRA) for Bloom-7b1
fit on the [Stanford-Alpaca-52k](https://github.com/tatsu-lab/stanford_alpaca)
and [databricks-dolly-15k](https://github.com/databrickslabs/dolly/tree/master/data) data in 52 languages.
### Dataset Creation
1. English Instructions: The English instuctions are obtained from [alpaca-52k](https://github.com/tatsu-lab/stanford_alpaca), and [dolly-15k](https://github.com/databrickslabs/dolly/tree/master/data).
2. Instruction Translation: The instructions (and inputs) are translated into the target languages using Google Translation API (conducted on April 2023).
3. Output Generation: We generate output from `gpt-3.5-turbo` for each language (conducted on April 2023).
<h3 align="center">
<img src="https://raw.githubusercontent.com/fajri91/eval_picts/master/BactrianX_dataset.jpg" width="950" align="center">
</h3>
### Training Parameters
The code for training the model is provided in our [github](https://github.com/mbzuai-nlp/Bactrian-X), which is adapted from [Alpaca-LoRA](https://github.com/tloen/alpaca-lora).
This version of the weights was trained with the following hyperparameters:
- Epochs: 10
- Batch size: 128
- Cutoff length: 512
- Learning rate: 3e-4
- Lora _r_: 64
- Lora target modules: query_key_value
That is:
```
python finetune.py \
--base_model='bigscience/bloom-7b1' \
--num_epochs=10 \
--batch_size=128 \
--cutoff_len=512 \
--group_by_length \
--output_dir='./bactrian-x-bloom-7b1-lora' \
--lora_target_modules='query_key_value' \
--lora_r=64 \
--micro_batch_size=32
```
Instructions for running it can be found at https://github.com/MBZUAI-nlp/Bactrian-X.
### Discussion of Biases
(1) Translation bias; (2) Potential English-culture bias in the translated dataset.
### Citation Information
```
@misc{li2023bactrianx,
title={Bactrian-X : A Multilingual Replicable Instruction-Following Model with Low-Rank Adaptation},
author={Haonan Li and Fajri Koto and Minghao Wu and Alham Fikri Aji and Timothy Baldwin},
year={2023},
eprint={2305.15011},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
| [
"TRANSLATION"
] | Non_BioNLP |
michaelfeil/ct2fast-opus-mt-ROMANCE-en | michaelfeil | translation | [
"transformers",
"ctranslate2",
"translation",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
] | 1,684,457,458,000 | 2023-05-19T00:51:47 | 352 | 1 | ---
license: apache-2.0
tags:
- ctranslate2
- translation
---
# # Fast-Inference with Ctranslate2
Speedup inference by 2x-8x using int8 inference in C++
quantized version of [Helsinki-NLP/opus-mt-ROMANCE-en](https://huggingface.co/Helsinki-NLP/opus-mt-ROMANCE-en)
```bash
pip install hf-hub-ctranslate2>=1.0.0 ctranslate2>=3.13.0
```
Converted using
```
ct2-transformers-converter --model Helsinki-NLP/opus-mt-ROMANCE-en --output_dir /home/michael/tmp-ct2fast-opus-mt-ROMANCE-en --force --copy_files README.md generation_config.json tokenizer_config.json vocab.json source.spm .gitattributes target.spm --quantization float16
```
Checkpoint compatible to [ctranslate2](https://github.com/OpenNMT/CTranslate2) and [hf-hub-ctranslate2](https://github.com/michaelfeil/hf-hub-ctranslate2)
- `compute_type=int8_float16` for `device="cuda"`
- `compute_type=int8` for `device="cpu"`
```python
from hf_hub_ctranslate2 import TranslatorCT2fromHfHub, GeneratorCT2fromHfHub
from transformers import AutoTokenizer
model_name = "michaelfeil/ct2fast-opus-mt-ROMANCE-en"
# use either TranslatorCT2fromHfHub or GeneratorCT2fromHfHub here, depending on model.
model = TranslatorCT2fromHfHub(
# load in int8 on CUDA
model_name_or_path=model_name,
device="cuda",
compute_type="int8_float16",
tokenizer=AutoTokenizer.from_pretrained("Helsinki-NLP/opus-mt-ROMANCE-en")
)
outputs = model.generate(
text=["How do you call a fast Flan-ingo?", "User: How are you doing?"],
)
print(outputs)
```
# Licence and other remarks:
This is just a quantized version. Licence conditions are intended to be idential to original huggingface repo.
# Original description
### opus-mt-ROMANCE-en
* source languages: fr,fr_BE,fr_CA,fr_FR,wa,frp,oc,ca,rm,lld,fur,lij,lmo,es,es_AR,es_CL,es_CO,es_CR,es_DO,es_EC,es_ES,es_GT,es_HN,es_MX,es_NI,es_PA,es_PE,es_PR,es_SV,es_UY,es_VE,pt,pt_br,pt_BR,pt_PT,gl,lad,an,mwl,it,it_IT,co,nap,scn,vec,sc,ro,la
* target languages: en
* OPUS readme: [fr+fr_BE+fr_CA+fr_FR+wa+frp+oc+ca+rm+lld+fur+lij+lmo+es+es_AR+es_CL+es_CO+es_CR+es_DO+es_EC+es_ES+es_GT+es_HN+es_MX+es_NI+es_PA+es_PE+es_PR+es_SV+es_UY+es_VE+pt+pt_br+pt_BR+pt_PT+gl+lad+an+mwl+it+it_IT+co+nap+scn+vec+sc+ro+la-en](https://github.com/Helsinki-NLP/OPUS-MT-train/blob/master/models/fr+fr_BE+fr_CA+fr_FR+wa+frp+oc+ca+rm+lld+fur+lij+lmo+es+es_AR+es_CL+es_CO+es_CR+es_DO+es_EC+es_ES+es_GT+es_HN+es_MX+es_NI+es_PA+es_PE+es_PR+es_SV+es_UY+es_VE+pt+pt_br+pt_BR+pt_PT+gl+lad+an+mwl+it+it_IT+co+nap+scn+vec+sc+ro+la-en/README.md)
* dataset: opus
* model: transformer
* pre-processing: normalization + SentencePiece
* download original weights: [opus-2020-04-01.zip](https://object.pouta.csc.fi/OPUS-MT-models/fr+fr_BE+fr_CA+fr_FR+wa+frp+oc+ca+rm+lld+fur+lij+lmo+es+es_AR+es_CL+es_CO+es_CR+es_DO+es_EC+es_ES+es_GT+es_HN+es_MX+es_NI+es_PA+es_PE+es_PR+es_SV+es_UY+es_VE+pt+pt_br+pt_BR+pt_PT+gl+lad+an+mwl+it+it_IT+co+nap+scn+vec+sc+ro+la-en/opus-2020-04-01.zip)
* test set translations: [opus-2020-04-01.test.txt](https://object.pouta.csc.fi/OPUS-MT-models/fr+fr_BE+fr_CA+fr_FR+wa+frp+oc+ca+rm+lld+fur+lij+lmo+es+es_AR+es_CL+es_CO+es_CR+es_DO+es_EC+es_ES+es_GT+es_HN+es_MX+es_NI+es_PA+es_PE+es_PR+es_SV+es_UY+es_VE+pt+pt_br+pt_BR+pt_PT+gl+lad+an+mwl+it+it_IT+co+nap+scn+vec+sc+ro+la-en/opus-2020-04-01.test.txt)
* test set scores: [opus-2020-04-01.eval.txt](https://object.pouta.csc.fi/OPUS-MT-models/fr+fr_BE+fr_CA+fr_FR+wa+frp+oc+ca+rm+lld+fur+lij+lmo+es+es_AR+es_CL+es_CO+es_CR+es_DO+es_EC+es_ES+es_GT+es_HN+es_MX+es_NI+es_PA+es_PE+es_PR+es_SV+es_UY+es_VE+pt+pt_br+pt_BR+pt_PT+gl+lad+an+mwl+it+it_IT+co+nap+scn+vec+sc+ro+la-en/opus-2020-04-01.eval.txt)
## Benchmarks
| testset | BLEU | chr-F |
|-----------------------|-------|-------|
| Tatoeba.fr.en | 62.2 | 0.750 |
| [
"TRANSLATION"
] | Non_BioNLP |
tsinik/distilbert-base-uncased-finetuned-emotion | tsinik | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,681,395,064,000 | 2023-04-14T06:26:43 | 13 | 0 | ---
datasets:
- emotion
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-emotion
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: emotion
type: emotion
args: split
metrics:
- type: accuracy
value: 0.9255
name: Accuracy
- type: f1
value: 0.9255660805721759
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-emotion
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2230
- Accuracy: 0.9255
- F1: 0.9256
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.8339 | 1.0 | 250 | 0.3241 | 0.9035 | 0.9006 |
| 0.2513 | 2.0 | 500 | 0.2230 | 0.9255 | 0.9256 |
### Framework versions
- Transformers 4.13.0
- Pytorch 2.0.0+cu118
- Datasets 2.8.0
- Tokenizers 0.10.3
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
seoultechLLM/Llama-3-70B-PIM-4bit | seoultechLLM | null | [
"license:mit",
"region:us"
] | 1,732,535,944,000 | 2024-11-25T12:05:09 | 0 | 0 | ---
license: mit
---
---
license: mit
---
# Model Architecture
## Base Model: Llama 3 (70 billion parameters)
## Quantization: 4-bit integer quantization for memory and computational efficiency
## Framework: Fine-tuned with PyTorch, leveraging Hugging Face Transformers
## PIM Optimization: Enhanced for PIM hardware to process data directly in memory, minimizing latency and maximizing throughput
# Intended Use
## Primary Use Cases:
Large-scale text generation
Summarization
Question answering
Conversational AI
Text classification
## Research Focus:
This model is specifically designed for research and industrial applications that require efficient handling of large language models with constrained hardware resources. | [
"TEXT_CLASSIFICATION",
"QUESTION_ANSWERING",
"SUMMARIZATION"
] | Non_BioNLP |
AIFS/Prometh-MOEM-24B | AIFS | text-generation | [
"transformers",
"safetensors",
"mixtral",
"text-generation",
"conversational",
"en",
"license:apache-2.0",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,707,832,848,000 | 2024-03-20T13:43:22 | 0 | 3 | ---
language:
- en
license: apache-2.0
---
# Prometh-MOEM-24B Model Card
**Prometh-MOEM-24B** is a Mixture of Experts (MoE) model that integrates multiple foundational models to deliver enhanced performance across a spectrum of tasks. It harnesses the combined strengths of its constituent models, optimizing for accuracy, speed, and versatility.
## Model Sources and Components
This MoE model incorporates the following specialized models:
- Language translation
- Question answering
## 💻Usage Instructions
```python
from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
# Load the model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("AIFS/Prometh-MOEM-24B")
model = AutoModelForCausalLM.from_pretrained("AIFS/Prometh-MOEM-24B")
# Set up the pipeline
text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer)
# Generate text
prompt = "The future of AI in healthcare is"
generated_texts = text_generator(prompt, max_length=50, num_return_sequences=3)
for generated_text in generated_texts:
print(generated_text["generated_text"])
```
## Technical Specifications
### Advanced Optimization
**Quantization and Fine-Tuning**: Prometh-MOEM-24B can be fine tuned, offering pathways for both quantization and fine-tuning. These processes refine the model's performance and efficiency, catering to the nuanced demands of deployment environments.
#### Quantization
Quantization is a technique aimed at reducing the computational and memory burdens of model inference. It achieves this feat by transitioning from high-precision data types, like 32-bit floating point (float32), to more compact and efficient formats, such as 8-bit integers (int8). This transition not only shrinks the model's memory footprint but also accelerates its operational pace, making it more viable for embedded systems or devices with limited computational resources.
- **Benefits**:
- **Application**:
- Prometh-MOEM-24B can be quantized post-training, adjusting to int8 without retraining from scratch. This method preserves the essence of its intelligence while adapting to the practical constraints of deployment environments.
#### Fine-Tuning
Beyond quantization, the model is primed for fine-tuning, allowing it to adapt to specific tasks or datasets with increased precision. This process involves additional training cycles on new data, thereby enhancing its acumen for particular applications.
- **Customization**: Tailors the model to specialized needs, optimizing its performance on tasks it was not originally designed for.
- **Versatility**: Ensures the model remains relevant and effective across a diverse array of use cases.
## Model Details and Attribution
- **Developed by:** [Iago Gaspar]
- **Shared by:** [AI Flow Solutions]
- **Model type:** Mixture of Experts Model
- **Language(s) (NLP):** en-en
- **License:** Apache-2.0
## Environmental Impact
## Out-of-Scope Use
The model is not intended for generating harmful or biased content.
## Bias, Risks, and Limitations
## Recommendations
Users should evaluate the model for biases and other ethical considerations before deploying it for real-world applications. | [
"QUESTION_ANSWERING",
"TRANSLATION"
] | Non_BioNLP |
RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf | RichardErkhov | null | [
"gguf",
"endpoints_compatible",
"region:us",
"conversational"
] | 1,731,793,513,000 | 2024-11-16T23:04:52 | 106 | 0 | ---
{}
---
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
gemma-ko-7b-instruct-v0.52 - GGUF
- Model creator: https://huggingface.co/lemon-mint/
- Original model: https://huggingface.co/lemon-mint/gemma-ko-7b-instruct-v0.52/
| Name | Quant method | Size |
| ---- | ---- | ---- |
| [gemma-ko-7b-instruct-v0.52.Q2_K.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q2_K.gguf) | Q2_K | 3.24GB |
| [gemma-ko-7b-instruct-v0.52.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q3_K_S.gguf) | Q3_K_S | 3.71GB |
| [gemma-ko-7b-instruct-v0.52.Q3_K.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q3_K.gguf) | Q3_K | 4.07GB |
| [gemma-ko-7b-instruct-v0.52.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q3_K_M.gguf) | Q3_K_M | 4.07GB |
| [gemma-ko-7b-instruct-v0.52.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q3_K_L.gguf) | Q3_K_L | 4.39GB |
| [gemma-ko-7b-instruct-v0.52.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.IQ4_XS.gguf) | IQ4_XS | 4.48GB |
| [gemma-ko-7b-instruct-v0.52.Q4_0.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q4_0.gguf) | Q4_0 | 4.67GB |
| [gemma-ko-7b-instruct-v0.52.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.IQ4_NL.gguf) | IQ4_NL | 4.69GB |
| [gemma-ko-7b-instruct-v0.52.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q4_K_S.gguf) | Q4_K_S | 4.7GB |
| [gemma-ko-7b-instruct-v0.52.Q4_K.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q4_K.gguf) | Q4_K | 4.96GB |
| [gemma-ko-7b-instruct-v0.52.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q4_K_M.gguf) | Q4_K_M | 4.96GB |
| [gemma-ko-7b-instruct-v0.52.Q4_1.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q4_1.gguf) | Q4_1 | 5.12GB |
| [gemma-ko-7b-instruct-v0.52.Q5_0.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q5_0.gguf) | Q5_0 | 5.57GB |
| [gemma-ko-7b-instruct-v0.52.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q5_K_S.gguf) | Q5_K_S | 5.57GB |
| [gemma-ko-7b-instruct-v0.52.Q5_K.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q5_K.gguf) | Q5_K | 5.72GB |
| [gemma-ko-7b-instruct-v0.52.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q5_K_M.gguf) | Q5_K_M | 5.72GB |
| [gemma-ko-7b-instruct-v0.52.Q5_1.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q5_1.gguf) | Q5_1 | 6.02GB |
| [gemma-ko-7b-instruct-v0.52.Q6_K.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q6_K.gguf) | Q6_K | 6.53GB |
| [gemma-ko-7b-instruct-v0.52.Q8_0.gguf](https://huggingface.co/RichardErkhov/lemon-mint_-_gemma-ko-7b-instruct-v0.52-gguf/blob/main/gemma-ko-7b-instruct-v0.52.Q8_0.gguf) | Q8_0 | 8.45GB |
Original model description:
---
library_name: transformers
license: other
license_name: gemma-terms-of-use
license_link: https://ai.google.dev/gemma/terms
language:
- ko
- en
tags:
- korean
- gemma
- pytorch
pipeline_tag: text-generation
base_model: beomi/gemma-ko-7b
---
# Gemma Ko 7B Instruct v0.52
- Eval Loss: `1.18263`
- Train Loss: `1.17373`
- lr: `1.7e-5`
- optimizer: adamw
- lr_scheduler_type: cosine
## Model Details
### Model Description
The Gemma Ko 7B Instruct v0.52 model is designed for generating human-like text in the Korean language.
It can be used for a variety of natural language processing tasks, such as language translation, text summarization, question answering, and conversation generation.
This model is particularly well-suited for applications that require high-quality, coherent, and contextually relevant Korean text generation.
- **Developed by:** `lemon-mint`
- **Model type:** Gemma
- **Language(s) (NLP):** Korean, English
- **License:** [gemma-terms-of-use](https://ai.google.dev/gemma/terms)
- **Finetuned from model:** [beomi/gemma-ko-7b](https://huggingface.co/beomi/gemma-ko-7b)
# Limitations and Ethical Considerations
As Gemma Ko 7B has been trained on extensive web data, biases present in the training data may be reflected in the model. Additionally, there is a possibility that it may generate sentences containing errors or incorrect information. Therefore, rather than blindly trusting the model's output, it is necessary to refer to it with caution.
| [
"QUESTION_ANSWERING",
"TRANSLATION",
"SUMMARIZATION"
] | Non_BioNLP |
DavieLion/Lllma-3.2-1B | DavieLion | text-generation | [
"transformers",
"safetensors",
"llama",
"text-generation",
"facebook",
"meta",
"pytorch",
"llama-3",
"en",
"de",
"fr",
"it",
"pt",
"hi",
"es",
"th",
"arxiv:2204.05149",
"arxiv:2405.16406",
"license:llama3.2",
"autotrain_compatible",
"text-generation-inference",
"endpoints_compatible",
"region:us"
] | 1,735,265,661,000 | 2024-12-27T07:19:30 | 15 | 0 | ---
language:
- en
- de
- fr
- it
- pt
- hi
- es
- th
library_name: transformers
license: llama3.2
pipeline_tag: text-generation
tags:
- facebook
- meta
- pytorch
- llama
- llama-3
extra_gated_prompt: "### LLAMA 3.2 COMMUNITY LICENSE AGREEMENT\n\nLlama 3.2 Version\
\ Release Date: September 25, 2024\n\n“Agreement” means the terms and conditions\
\ for use, reproduction, distribution and modification of the Llama Materials set\
\ forth herein.\n\n“Documentation” means the specifications, manuals and documentation\
\ accompanying Llama 3.2 distributed by Meta at https://llama.meta.com/doc/overview.\n\
\n“Licensee” or “you” means you, or your employer or any other person or entity\
\ (if you are entering into this Agreement on such person or entity’s behalf),\
\ of the age required under applicable laws, rules or regulations to provide legal\
\ consent and that has legal authority to bind your employer or such other person\
\ or entity if you are entering in this Agreement on their behalf.\n\n“Llama 3.2”\
\ means the foundational large language models and software and algorithms, including\
\ machine-learning model code, trained model weights, inference-enabling code, training-enabling\
\ code, fine-tuning enabling code and other elements of the foregoing distributed\
\ by Meta at https://www.llama.com/llama-downloads.\n\n“Llama Materials” means,\
\ collectively, Meta’s proprietary Llama 3.2 and Documentation (and any portion\
\ thereof) made available under this Agreement.\n\n“Meta” or “we” means Meta Platforms\
\ Ireland Limited (if you are located in or, if you are an entity, your principal\
\ place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if\
\ you are located outside of the EEA or Switzerland). \n\nBy clicking “I Accept”\
\ below or by using or distributing any portion or element of the Llama Materials,\
\ you agree to be bound by this Agreement.\n\n1. License Rights and Redistribution.\n\
a. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable\
\ and royalty-free limited license under Meta’s intellectual property or other rights\
\ owned by Meta embodied in the Llama Materials to use, reproduce, distribute,\
\ copy, create derivative works of, and make modifications to the Llama Materials.\
\ \nb. Redistribution and Use. \ni. If you distribute or make available the Llama\
\ Materials (or any derivative works thereof), or a product or service (including\
\ another AI model) that contains any of them, you shall (A) provide a copy of this\
\ Agreement with any such Llama Materials; and (B) prominently display “Built with\
\ Llama” on a related website, user interface, blogpost, about page, or product\
\ documentation. If you use the Llama Materials or any outputs or results of the\
\ Llama Materials to create, train, fine tune, or otherwise improve an AI model,\
\ which is distributed or made available, you shall also include “Llama” at the\
\ beginning of any such AI model name.\nii. If you receive Llama Materials, or any\
\ derivative works thereof, from a Licensee as part of an integrated end user product,\
\ then Section 2 of this Agreement will not apply to you. \niii. You must retain\
\ in all copies of the Llama Materials that you distribute the following attribution\
\ notice within a “Notice” text file distributed as a part of such copies: “Llama\
\ 3.2 is licensed under the Llama 3.2 Community License, Copyright © Meta Platforms,\
\ Inc. All Rights Reserved.”\niv. Your use of the Llama Materials must comply with\
\ applicable laws and regulations (including trade compliance laws and regulations)\
\ and adhere to the Acceptable Use Policy for the Llama Materials (available at\
\ https://www.llama.com/llama3_2/use-policy), which is hereby incorporated by reference\
\ into this Agreement.\n \n2. Additional Commercial Terms. If, on the Llama 3.2\
\ version release date, the monthly active users of the products or services made\
\ available by or for Licensee, or Licensee’s affiliates, is greater than 700 million\
\ monthly active users in the preceding calendar month, you must request a license\
\ from Meta, which Meta may grant to you in its sole discretion, and you are not\
\ authorized to exercise any of the rights under this Agreement unless or until\
\ Meta otherwise expressly grants you such rights.\n3. Disclaimer of Warranty. UNLESS\
\ REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM\
\ ARE PROVIDED ON AN “AS IS” BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS\
\ ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION,\
\ ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR\
\ PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING\
\ OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR\
\ USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS.\n4. Limitation of Liability.\
\ IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY,\
\ WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING\
\ OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL,\
\ INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE\
\ BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING.\n5. Intellectual Property.\n\
a. No trademark licenses are granted under this Agreement, and in connection with\
\ the Llama Materials, neither Meta nor Licensee may use any name or mark owned\
\ by or associated with the other or any of its affiliates, except as required\
\ for reasonable and customary use in describing and redistributing the Llama Materials\
\ or as set forth in this Section 5(a). Meta hereby grants you a license to use\
\ “Llama” (the “Mark”) solely as required to comply with the last sentence of Section\
\ 1.b.i. You will comply with Meta’s brand guidelines (currently accessible at\
\ https://about.meta.com/brand/resources/meta/company-brand/). All goodwill arising\
\ out of your use of the Mark will inure to the benefit of Meta.\nb. Subject to\
\ Meta’s ownership of Llama Materials and derivatives made by or for Meta, with\
\ respect to any derivative works and modifications of the Llama Materials that\
\ are made by you, as between you and Meta, you are and will be the owner of such\
\ derivative works and modifications.\nc. If you institute litigation or other proceedings\
\ against Meta or any entity (including a cross-claim or counterclaim in a lawsuit)\
\ alleging that the Llama Materials or Llama 3.2 outputs or results, or any portion\
\ of any of the foregoing, constitutes infringement of intellectual property or\
\ other rights owned or licensable by you, then any licenses granted to you under\
\ this Agreement shall terminate as of the date such litigation or claim is filed\
\ or instituted. You will indemnify and hold harmless Meta from and against any\
\ claim by any third party arising out of or related to your use or distribution\
\ of the Llama Materials.\n6. Term and Termination. The term of this Agreement will\
\ commence upon your acceptance of this Agreement or access to the Llama Materials\
\ and will continue in full force and effect until terminated in accordance with\
\ the terms and conditions herein. Meta may terminate this Agreement if you are\
\ in breach of any term or condition of this Agreement. Upon termination of this\
\ Agreement, you shall delete and cease use of the Llama Materials. Sections 3,\
\ 4 and 7 shall survive the termination of this Agreement. \n7. Governing Law and\
\ Jurisdiction. This Agreement will be governed and construed under the laws of\
\ the State of California without regard to choice of law principles, and the UN\
\ Convention on Contracts for the International Sale of Goods does not apply to\
\ this Agreement. The courts of California shall have exclusive jurisdiction of\
\ any dispute arising out of this Agreement. \n### Llama 3.2 Acceptable Use Policy\n\
Meta is committed to promoting safe and fair use of its tools and features, including\
\ Llama 3.2. If you access or use Llama 3.2, you agree to this Acceptable Use Policy\
\ (“**Policy**”). The most recent copy of this policy can be found at [https://www.llama.com/llama3_2/use-policy](https://www.llama.com/llama3_2/use-policy).\n\
#### Prohibited Uses\nWe want everyone to use Llama 3.2 safely and responsibly.\
\ You agree you will not use, or allow others to use, Llama 3.2 to:\n1. Violate\
\ the law or others’ rights, including to:\n 1. Engage in, promote, generate,\
\ contribute to, encourage, plan, incite, or further illegal or unlawful activity\
\ or content, such as:\n 1. Violence or terrorism\n 2. Exploitation\
\ or harm to children, including the solicitation, creation, acquisition, or dissemination\
\ of child exploitative content or failure to report Child Sexual Abuse Material\n\
\ 3. Human trafficking, exploitation, and sexual violence\n 4. The\
\ illegal distribution of information or materials to minors, including obscene\
\ materials, or failure to employ legally required age-gating in connection with\
\ such information or materials.\n 5. Sexual solicitation\n 6. Any\
\ other criminal activity\n 1. Engage in, promote, incite, or facilitate the\
\ harassment, abuse, threatening, or bullying of individuals or groups of individuals\n\
\ 2. Engage in, promote, incite, or facilitate discrimination or other unlawful\
\ or harmful conduct in the provision of employment, employment benefits, credit,\
\ housing, other economic benefits, or other essential goods and services\n 3.\
\ Engage in the unauthorized or unlicensed practice of any profession including,\
\ but not limited to, financial, legal, medical/health, or related professional\
\ practices\n 4. Collect, process, disclose, generate, or infer private or sensitive\
\ information about individuals, including information about individuals’ identity,\
\ health, or demographic information, unless you have obtained the right to do so\
\ in accordance with applicable law\n 5. Engage in or facilitate any action or\
\ generate any content that infringes, misappropriates, or otherwise violates any\
\ third-party rights, including the outputs or results of any products or services\
\ using the Llama Materials\n 6. Create, generate, or facilitate the creation\
\ of malicious code, malware, computer viruses or do anything else that could disable,\
\ overburden, interfere with or impair the proper working, integrity, operation\
\ or appearance of a website or computer system\n 7. Engage in any action, or\
\ facilitate any action, to intentionally circumvent or remove usage restrictions\
\ or other safety measures, or to enable functionality disabled by Meta \n2. Engage\
\ in, promote, incite, facilitate, or assist in the planning or development of activities\
\ that present a risk of death or bodily harm to individuals, including use of Llama\
\ 3.2 related to the following:\n 8. Military, warfare, nuclear industries or\
\ applications, espionage, use for materials or activities that are subject to the\
\ International Traffic Arms Regulations (ITAR) maintained by the United States\
\ Department of State or to the U.S. Biological Weapons Anti-Terrorism Act of 1989\
\ or the Chemical Weapons Convention Implementation Act of 1997\n 9. Guns and\
\ illegal weapons (including weapon development)\n 10. Illegal drugs and regulated/controlled\
\ substances\n 11. Operation of critical infrastructure, transportation technologies,\
\ or heavy machinery\n 12. Self-harm or harm to others, including suicide, cutting,\
\ and eating disorders\n 13. Any content intended to incite or promote violence,\
\ abuse, or any infliction of bodily harm to an individual\n3. Intentionally deceive\
\ or mislead others, including use of Llama 3.2 related to the following:\n 14.\
\ Generating, promoting, or furthering fraud or the creation or promotion of disinformation\n\
\ 15. Generating, promoting, or furthering defamatory content, including the\
\ creation of defamatory statements, images, or other content\n 16. Generating,\
\ promoting, or further distributing spam\n 17. Impersonating another individual\
\ without consent, authorization, or legal right\n 18. Representing that the\
\ use of Llama 3.2 or outputs are human-generated\n 19. Generating or facilitating\
\ false online engagement, including fake reviews and other means of fake online\
\ engagement \n4. Fail to appropriately disclose to end users any known dangers\
\ of your AI system 5. Interact with third party tools, models, or software designed\
\ to generate unlawful content or engage in unlawful or harmful conduct and/or represent\
\ that the outputs of such tools, models, or software are associated with Meta or\
\ Llama 3.2\n\nWith respect to any multimodal models included in Llama 3.2, the\
\ rights granted under Section 1(a) of the Llama 3.2 Community License Agreement\
\ are not being granted to you if you are an individual domiciled in, or a company\
\ with a principal place of business in, the European Union. This restriction does\
\ not apply to end users of a product or service that incorporates any such multimodal\
\ models.\n\nPlease report any violation of this Policy, software “bug,” or other\
\ problems that could lead to a violation of this Policy through one of the following\
\ means:\n\n* Reporting issues with the model: [https://github.com/meta-llama/llama-models/issues](https://l.workplace.com/l.php?u=https%3A%2F%2Fgithub.com%2Fmeta-llama%2Fllama-models%2Fissues&h=AT0qV8W9BFT6NwihiOHRuKYQM_UnkzN_NmHMy91OT55gkLpgi4kQupHUl0ssR4dQsIQ8n3tfd0vtkobvsEvt1l4Ic6GXI2EeuHV8N08OG2WnbAmm0FL4ObkazC6G_256vN0lN9DsykCvCqGZ)\n\
* Reporting risky content generated by the model: [developers.facebook.com/llama_output_feedback](http://developers.facebook.com/llama_output_feedback)\n\
* Reporting bugs and security concerns: [facebook.com/whitehat/info](http://facebook.com/whitehat/info)\n\
* Reporting violations of the Acceptable Use Policy or unlicensed uses of Llama\
\ 3.2: [email protected]"
extra_gated_fields:
First Name: text
Last Name: text
Date of birth: date_picker
Country: country
Affiliation: text
Job title:
type: select
options:
- Student
- Research Graduate
- AI researcher
- AI developer/engineer
- Reporter
- Other
geo: ip_location
? By clicking Submit below I accept the terms of the license and acknowledge that
the information I provide will be collected stored processed and shared in accordance
with the Meta Privacy Policy
: checkbox
extra_gated_description: The information you provide will be collected, stored, processed
and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/).
extra_gated_button_content: Submit
---
## Model Information
The Llama 3.2 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction-tuned generative models in 1B and 3B sizes (text in/text out). The Llama 3.2 instruction-tuned text only models are optimized for multilingual dialogue use cases, including agentic retrieval and summarization tasks. They outperform many of the available open source and closed chat models on common industry benchmarks.
**Model Developer:** Meta
**Model Architecture:** Llama 3.2 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
| | Training Data | Params | Input modalities | Output modalities | Context Length | GQA | Shared Embeddings | Token count | Knowledge cutoff |
| :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- |
| Llama 3.2 (text only) | A new mix of publicly available online data. | 1B (1.23B) | Multilingual Text | Multilingual Text and code | 128k | Yes | Yes | Up to 9T tokens | December 2023 |
| | | 3B (3.21B) | Multilingual Text | Multilingual Text and code | | | | | |
| Llama 3.2 Quantized (text only) | A new mix of publicly available online data. | 1B (1.23B) | Multilingual Text | Multilingual Text and code | 8k | Yes | Yes | Up to 9T tokens | December 2023 |
| | | 3B (3.21B) | Multilingual Text | Multilingual Text and code | | | | | |
**Supported Languages:** English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai are officially supported. Llama 3.2 has been trained on a broader collection of languages than these 8 supported languages. Developers may fine-tune Llama 3.2 models for languages beyond these supported languages, provided they comply with the Llama 3.2 Community License and the Acceptable Use Policy. Developers are always expected to ensure that their deployments, including those that involve additional languages, are completed safely and responsibly.
**Llama 3.2 Model Family:** Token counts refer to pretraining data only. All model versions use Grouped-Query Attention (GQA) for improved inference scalability.
**Model Release Date:** Sept 25, 2024
**Status:** This is a static model trained on an offline dataset. Future versions may be released that improve model capabilities and safety.
**License:** Use of Llama 3.2 is governed by the [Llama 3.2 Community License](https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/LICENSE) (a custom, commercial license agreement).
**Feedback:** Instructions on how to provide feedback or comments on the model can be found in the Llama Models [README](https://github.com/meta-llama/llama-models/blob/main/README.md). For more technical information about generation parameters and recipes for how to use Llama 3.2 in applications, please go [here](https://github.com/meta-llama/llama-recipes).
## Intended Use
**Intended Use Cases:** Llama 3.2 is intended for commercial and research use in multiple languages. Instruction tuned text only models are intended for assistant-like chat and agentic applications like knowledge retrieval and summarization, mobile AI powered writing assistants and query and prompt rewriting. Pretrained models can be adapted for a variety of additional natural language generation tasks. Similarly, quantized models can be adapted for a variety of on-device use-cases with limited compute resources.
**Out of Scope:** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3.2 Community License. Use in languages beyond those explicitly referenced as supported in this model card.
## How to use
This repository contains two versions of Llama-3.2-1B, for use with transformers and with the original `llama` codebase.
### Use with transformers
Starting with transformers >= 4.43.0 onward, you can run conversational inference using the Transformers pipeline abstraction or by leveraging the Auto classes with the generate() function.
Make sure to update your transformers installation via pip install --upgrade transformers.
```python
import torch
from transformers import pipeline
model_id = "meta-llama/Llama-3.2-1B"
pipe = pipeline(
"text-generation",
model=model_id,
torch_dtype=torch.bfloat16,
device_map="auto"
)
pipe("The key to life is")
```
### Use with `llama`
Please, follow the instructions in the [repository](https://github.com/meta-llama/llama).
To download Original checkpoints, see the example command below leveraging `huggingface-cli`:
```
huggingface-cli download meta-llama/Llama-3.2-1B --include "original/*" --local-dir Llama-3.2-1B
```
## Hardware and Software
**Training Factors:** We used custom training libraries, Meta's custom built GPU cluster, and production infrastructure for pretraining. Fine-tuning, quantization, annotation, and evaluation were also performed on production infrastructure.
**Training Energy Use:** Training utilized a cumulative of **916k** GPU hours of computation on H100-80GB (TDP of 700W) type hardware, per the table below. Training time is the total GPU time required for training each model and power consumption is the peak power capacity per GPU device used, adjusted for power usage efficiency.
**Training Greenhouse Gas Emissions:** Estimated total location-based greenhouse gas emissions were **240** tons CO2eq for training. Since 2020, Meta has maintained net zero greenhouse gas emissions in its global operations and matched 100% of its electricity use with renewable energy; therefore, the total market-based greenhouse gas emissions for training were 0 tons CO2eq.
| | Training Time (GPU hours) | Logit Generation Time (GPU Hours) | Training Power Consumption (W) | Training Location-Based Greenhouse Gas Emissions (tons CO2eq) | Training Market-Based Greenhouse Gas Emissions (tons CO2eq) |
| :---- | :---: | ----- | :---: | :---: | :---: |
| Llama 3.2 1B | 370k | \- | 700 | 107 | 0 |
| Llama 3.2 3B | 460k | \- | 700 | 133 | 0 |
| Llama 3.2 1B SpinQuant | 1.7 | 0 | 700 | *Negligible*\*\* | 0 |
| Llama 3.2 3B SpinQuant | 2.4 | 0 | 700 | *Negligible*\*\* | 0 |
| Llama 3.2 1B QLora | 1.3k | 0 | 700 | 0.381 | 0 |
| Llama 3.2 3B QLora | 1.6k | 0 | 700 | 0.461 | 0 |
| Total | 833k | 86k | | 240 | 0 |
\*\* The location-based CO2e emissions of Llama 3.2 1B SpinQuant and Llama 3.2 3B SpinQuant are less than 0.001 metric tonnes each. This is due to the minimal training GPU hours that are required.
The methodology used to determine training energy use and greenhouse gas emissions can be found [here](https://arxiv.org/pdf/2204.05149). Since Meta is openly releasing these models, the training energy use and greenhouse gas emissions will not be incurred by others.
## Training Data
**Overview:** Llama 3.2 was pretrained on up to 9 trillion tokens of data from publicly available sources. For the 1B and 3B Llama 3.2 models, we incorporated logits from the Llama 3.1 8B and 70B models into the pretraining stage of the model development, where outputs (logits) from these larger models were used as token-level targets. Knowledge distillation was used after pruning to recover performance. In post-training we used a similar recipe as Llama 3.1 and produced final chat models by doing several rounds of alignment on top of the pre-trained model. Each round involved Supervised Fine-Tuning (SFT), Rejection Sampling (RS), and Direct Preference Optimization (DPO).
**Data Freshness:** The pretraining data has a cutoff of December 2023\.
## Quantization
### Quantization Scheme
We designed the current quantization scheme with the [PyTorch’s ExecuTorch](https://github.com/pytorch/executorch) inference framework and Arm CPU backend in mind, taking into account metrics including model quality, prefill/decoding speed, and memory footprint. Our quantization scheme involves three parts:
- All linear layers in all transformer blocks are quantized to a 4-bit groupwise scheme (with a group size of 32) for weights and 8-bit per-token dynamic quantization for activations.
- The classification layer is quantized to 8-bit per-channel for weight and 8-bit per token dynamic quantization for activation.
- Similar to classification layer, an 8-bit per channel quantization is used for embedding layer.
### Quantization-Aware Training and LoRA
The quantization-aware training (QAT) with low-rank adaptation (LoRA) models went through only post-training stages, using the same data as the full precision models. To initialize QAT, we utilize BF16 Llama 3.2 model checkpoints obtained after supervised fine-tuning (SFT) and perform an additional full round of SFT training with QAT. We then freeze the backbone of the QAT model and perform another round of SFT with LoRA adaptors applied to all layers within the transformer block. Meanwhile, the LoRA adaptors' weights and activations are maintained in BF16. Because our approach is similar to QLoRA of Dettmers et al., (2023) (i.e., quantization followed by LoRA adapters), we refer this method as QLoRA. Finally, we fine-tune the resulting model (both backbone and LoRA adaptors) using direct preference optimization (DPO).
### SpinQuant
[SpinQuant](https://arxiv.org/abs/2405.16406) was applied, together with generative post-training quantization (GPTQ). For the SpinQuant rotation matrix fine-tuning, we optimized for 100 iterations, using 800 samples with sequence-length 2048 from the WikiText 2 dataset. For GPTQ, we used 128 samples from the same dataset with the same sequence-length.
## Benchmarks \- English Text
In this section, we report the results for Llama 3.2 models on standard automatic benchmarks. For all these evaluations, we used our internal evaluations library.
### Base Pretrained Models
| Category | Benchmark | \# Shots | Metric | Llama 3.2 1B | Llama 3.2 3B | Llama 3.1 8B |
| ----- | ----- | :---: | :---: | :---: | :---: | :---: |
| General | MMLU | 5 | macro\_avg/acc\_char | 32.2 | 58 | 66.7 |
| | AGIEval English | 3-5 | average/acc\_char | 23.3 | 39.2 | 47.8 |
| | ARC-Challenge | 25 | acc\_char | 32.8 | 69.1 | 79.7 |
| Reading comprehension | SQuAD | 1 | em | 49.2 | 67.7 | 77 |
| | QuAC (F1) | 1 | f1 | 37.9 | 42.9 | 44.9 |
| | DROP (F1) | 3 | f1 | 28.0 | 45.2 | 59.5 |
| Long Context | Needle in Haystack | 0 | em | 96.8 | 1 | 1 |
### Instruction Tuned Models
| Capability | | Benchmark | \# Shots | Metric | Llama 3.2 1B bf16 | Llama 3.2 1B Vanilla PTQ\*\* | Llama 3.2 1B Spin Quant | Llama 3.2 1B QLoRA | Llama 3.2 3B bf16 | Llama 3.2 3B Vanilla PTQ\*\* | Llama 3.2 3B Spin Quant | Llama 3.2 3B QLoRA | Llama 3.1 8B |
| :---: | ----- | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| General | | MMLU | 5 | macro\_avg/acc | 49.3 | 43.3 | 47.3 | 49.0 | 63.4 | 60.5 | 62 | 62.4 | 69.4 |
| Re-writing | | Open-rewrite eval | 0 | micro\_avg/rougeL | 41.6 | 39.2 | 40.9 | 41.2 | 40.1 | 40.3 | 40.8 | 40.7 | 40.9 |
| Summarization | | TLDR9+ (test) | 1 | rougeL | 16.8 | 14.9 | 16.7 | 16.8 | 19.0 | 19.1 | 19.2 | 19.1 | 17.2 |
| Instruction following | | IFEval | 0 | Avg(Prompt/Instruction acc Loose/Strict) | 59.5 | 51.5 | 58.4 | 55.6 | 77.4 | 73.9 | 73.5 | 75.9 | 80.4 |
| Math | | GSM8K (CoT) | 8 | em\_maj1@1 | 44.4 | 33.1 | 40.6 | 46.5 | 77.7 | 72.9 | 75.7 | 77.9 | 84.5 |
| | | MATH (CoT) | 0 | final\_em | 30.6 | 20.5 | 25.3 | 31.0 | 48.0 | 44.2 | 45.3 | 49.2 | 51.9 |
| Reasoning | | ARC-C | 0 | acc | 59.4 | 54.3 | 57 | 60.7 | 78.6 | 75.6 | 77.6 | 77.6 | 83.4 |
| | | GPQA | 0 | acc | 27.2 | 25.9 | 26.3 | 25.9 | 32.8 | 32.8 | 31.7 | 33.9 | 32.8 |
| | | Hellaswag | 0 | acc | 41.2 | 38.1 | 41.3 | 41.5 | 69.8 | 66.3 | 68 | 66.3 | 78.7 |
| Tool Use | | BFCL V2 | 0 | acc | 25.7 | 14.3 | 15.9 | 23.7 | 67.0 | 53.4 | 60.1 | 63.5 | 67.1 |
| | | Nexus | 0 | macro\_avg/acc | 13.5 | 5.2 | 9.6 | 12.5 | 34.3 | 32.4 | 31.5 | 30.1 | 38.5 |
| Long Context | | InfiniteBench/En.QA | 0 | longbook\_qa/f1 | 20.3 | N/A | N/A | N/A | 19.8 | N/A | N/A | N/A | 27.3 |
| | | InfiniteBench/En.MC | 0 | longbook\_choice/acc | 38.0 | N/A | N/A | N/A | 63.3 | N/A | N/A | N/A | 72.2 |
| | | NIH/Multi-needle | 0 | recall | 75.0 | N/A | N/A | N/A | 84.7 | N/A | N/A | N/A | 98.8 |
| Multilingual | | MGSM (CoT) | 0 | em | 24.5 | 13.7 | 18.2 | 24.4 | 58.2 | 48.9 | 54.3 | 56.8 | 68.9 |
\*\*for comparison purposes only. Model not released.
### Multilingual Benchmarks
| Category | Benchmark | Language | Llama 3.2 1B | Llama 3.2 1B Vanilla PTQ\*\* | Llama 3.2 1B Spin Quant | Llama 3.2 1B QLoRA | Llama 3.2 3B | Llama 3.2 3B Vanilla PTQ\*\* | Llama 3.2 3B Spin Quant | Llama 3.2 3B QLoRA | Llama 3.1 8B |
| :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| General | MMLU (5-shot, macro_avg/acc) | Portuguese | 39.8 | 34.9 | 38.9 | 40.2 | 54.5 | 50.9 | 53.3 | 53.4 | 62.1 |
| | | Spanish | 41.5 | 36.0 | 39.8 | 41.8 | 55.1 | 51.9 | 53.6 | 53.6 | 62.5 |
| | | Italian | 39.8 | 34.9 | 38.1 | 40.6 | 53.8 | 49.9 | 52.1 | 51.7 | 61.6 |
| | | German | 39.2 | 34.9 | 37.5 | 39.6 | 53.3 | 50.0 | 52.2 | 51.3 | 60.6 |
| | | French | 40.5 | 34.8 | 39.2 | 40.8 | 54.6 | 51.2 | 53.3 | 53.3 | 62.3 |
| | | Hindi | 33.5 | 30.0 | 32.1 | 34.0 | 43.3 | 40.4 | 42.0 | 42.1 | 50.9 |
| | | Thai | 34.7 | 31.2 | 32.4 | 34.9 | 44.5 | 41.3 | 44.0 | 42.2 | 50.3 |
\*\*for comparison purposes only. Model not released.
## Inference time
In the below table, we compare the performance metrics of different quantization methods (SpinQuant and QAT \+ LoRA) with the BF16 baseline. The evaluation was done using the [ExecuTorch](https://github.com/pytorch/executorch) framework as the inference engine, with the ARM CPU as a backend using Android OnePlus 12 device.
| Category | Decode (tokens/sec) | Time-to-first-token (sec) | Prefill (tokens/sec) | Model size (PTE file size in MB) | Memory size (RSS in MB) |
| :---- | ----- | ----- | ----- | ----- | ----- |
| 1B BF16 (baseline) | 19.2 | 1.0 | 60.3 | 2358 | 3,185 |
| 1B SpinQuant | 50.2 (2.6x) | 0.3 (-76.9%) | 260.5 (4.3x) | 1083 (-54.1%) | 1,921 (-39.7%) |
| 1B QLoRA | 45.8 (2.4x) | 0.3 (-76.0%) | 252.0 (4.2x) | 1127 (-52.2%) | 2,255 (-29.2%) |
| 3B BF16 (baseline) | 7.6 | 3.0 | 21.2 | 6129 | 7,419 |
| 3B SpinQuant | 19.7 (2.6x) | 0.7 (-76.4%) | 89.7 (4.2x) | 2435 (-60.3%) | 3,726 (-49.8%) |
| 3B QLoRA | 18.5 (2.4x) | 0.7 (-76.1%) | 88.8 (4.2x) | 2529 (-58.7%) | 4,060 (-45.3%) |
(\*) The performance measurement is done using an adb binary-based approach.
(\*\*) It is measured on an Android OnePlus 12 device.
(\*\*\*) Time-to-first-token (TTFT) is measured with prompt length=64
*Footnote:*
- *Decode (tokens/second) is for how quickly it keeps generating. Higher is better.*
- *Time-to-first-token (TTFT for shorthand) is for how fast it generates the first token for a given prompt. Lower is better.*
- *Prefill is the inverse of TTFT (aka 1/TTFT) in tokens/second. Higher is better*
- *Model size \- how big is the model, measured by, PTE file, a binary file format for ExecuTorch*
- *RSS size \- Memory usage in resident set size (RSS)*
## Responsibility & Safety
As part of our Responsible release approach, we followed a three-pronged strategy to managing trust & safety risks:
1. Enable developers to deploy helpful, safe and flexible experiences for their target audience and for the use cases supported by Llama
2. Protect developers against adversarial users aiming to exploit Llama capabilities to potentially cause harm
3. Provide protections for the community to help prevent the misuse of our models
### Responsible Deployment
**Approach:** Llama is a foundational technology designed to be used in a variety of use cases. Examples on how Meta’s Llama models have been responsibly deployed can be found in our [Community Stories webpage](https://llama.meta.com/community-stories/). Our approach is to build the most helpful models, enabling the world to benefit from the technology power, by aligning our model safety for generic use cases and addressing a standard set of harms. Developers are then in the driver’s seat to tailor safety for their use cases, defining their own policies and deploying the models with the necessary safeguards in their Llama systems. Llama 3.2 was developed following the best practices outlined in our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/).
#### Llama 3.2 Instruct
**Objective:** Our main objectives for conducting safety fine-tuning are to provide the research community with a valuable resource for studying the robustness of safety fine-tuning, as well as to offer developers a readily available, safe, and powerful model for various applications to reduce the developer workload to deploy safe AI systems. We implemented the same set of safety mitigations as in Llama 3, and you can learn more about these in the Llama 3 [paper](https://ai.meta.com/research/publications/the-llama-3-herd-of-models/).
**Fine-Tuning Data:** We employ a multi-faceted approach to data collection, combining human-generated data from our vendors with synthetic data to mitigate potential safety risks. We’ve developed many large language model (LLM)-based classifiers that enable us to thoughtfully select high-quality prompts and responses, enhancing data quality control.
**Refusals and Tone:** Building on the work we started with Llama 3, we put a great emphasis on model refusals to benign prompts as well as refusal tone. We included both borderline and adversarial prompts in our safety data strategy, and modified our safety data responses to follow tone guidelines.
#### Llama 3.2 Systems
**Safety as a System:** Large language models, including Llama 3.2, **are not designed to be deployed in isolation** but instead should be deployed as part of an overall AI system with additional safety guardrails as required. Developers are expected to deploy system safeguards when building agentic systems. Safeguards are key to achieve the right helpfulness-safety alignment as well as mitigating safety and security risks inherent to the system and any integration of the model or system with external tools. As part of our responsible release approach, we provide the community with [safeguards](https://llama.meta.com/trust-and-safety/) that developers should deploy with Llama models or other LLMs, including Llama Guard, Prompt Guard and Code Shield. All our [reference implementations](https://github.com/meta-llama/llama-agentic-system) demos contain these safeguards by default so developers can benefit from system-level safety out-of-the-box.
### New Capabilities and Use Cases
**Technological Advancement:** Llama releases usually introduce new capabilities that require specific considerations in addition to the best practices that generally apply across all Generative AI use cases. For prior release capabilities also supported by Llama 3.2, see [Llama 3.1 Model Card](https://github.com/meta-llama/llama-models/blob/main/models/llama3_1/MODEL_CARD.md), as the same considerations apply here as well.
**Constrained Environments:** Llama 3.2 1B and 3B models are expected to be deployed in highly constrained environments, such as mobile devices. LLM Systems using smaller models will have a different alignment profile and safety/helpfulness tradeoff than more complex, larger systems. Developers should ensure the safety of their system meets the requirements of their use case. We recommend using lighter system safeguards for such use cases, like Llama Guard 3-1B or its mobile-optimized version.
### Evaluations
**Scaled Evaluations:** We built dedicated, adversarial evaluation datasets and evaluated systems composed of Llama models and Purple Llama safeguards to filter input prompt and output response. It is important to evaluate applications in context, and we recommend building dedicated evaluation dataset for your use case.
**Red Teaming:** We conducted recurring red teaming exercises with the goal of discovering risks via adversarial prompting and we used the learnings to improve our benchmarks and safety tuning datasets. We partnered early with subject-matter experts in critical risk areas to understand the nature of these real-world harms and how such models may lead to unintended harm for society. Based on these conversations, we derived a set of adversarial goals for the red team to attempt to achieve, such as extracting harmful information or reprogramming the model to act in a potentially harmful capacity. The red team consisted of experts in cybersecurity, adversarial machine learning, responsible AI, and integrity in addition to multilingual content specialists with background in integrity issues in specific geographic markets.
### Critical Risks
In addition to our safety work above, we took extra care on measuring and/or mitigating the following critical risk areas:
**1\. CBRNE (Chemical, Biological, Radiological, Nuclear, and Explosive Weapons):** Llama 3.2 1B and 3B models are smaller and less capable derivatives of Llama 3.1. For Llama 3.1 70B and 405B, to assess risks related to proliferation of chemical and biological weapons, we performed uplift testing designed to assess whether use of Llama 3.1 models could meaningfully increase the capabilities of malicious actors to plan or carry out attacks using these types of weapons and have determined that such testing also applies to the smaller 1B and 3B models.
**2\. Child Safety:** Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors including the additional languages Llama 3 is trained on. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences.
**3\. Cyber Attacks:** For Llama 3.1 405B, our cyber attack uplift study investigated whether LLMs can enhance human capabilities in hacking tasks, both in terms of skill level and speed.
Our attack automation study focused on evaluating the capabilities of LLMs when used as autonomous agents in cyber offensive operations, specifically in the context of ransomware attacks. This evaluation was distinct from previous studies that considered LLMs as interactive assistants. The primary objective was to assess whether these models could effectively function as independent agents in executing complex cyber-attacks without human intervention. Because Llama 3.2’s 1B and 3B models are smaller and less capable models than Llama 3.1 405B, we broadly believe that the testing conducted for the 405B model also applies to Llama 3.2 models.
### Community
**Industry Partnerships:** Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership on AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama).
**Grants:** We also set up the [Llama Impact Grants](https://llama.meta.com/llama-impact-grants/) program to identify and support the most compelling applications of Meta’s Llama model for societal benefit across three categories: education, climate and open innovation. The 20 finalists from the hundreds of applications can be found [here](https://llama.meta.com/llama-impact-grants/#finalists).
**Reporting:** Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community.
## Ethical Considerations and Limitations
**Values:** The core values of Llama 3.2 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3.2 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress.
**Testing:** Llama 3.2 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3.2’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3.2 models, developers should perform safety testing and tuning tailored to their specific applications of the model. Please refer to available resources including our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide), [Trust and Safety](https://llama.meta.com/trust-and-safety/) solutions, and other [resources](https://llama.meta.com/docs/get-started/) to learn more about responsible development.
| [
"SUMMARIZATION"
] | Non_BioNLP |
RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf | RichardErkhov | null | [
"gguf",
"endpoints_compatible",
"region:us"
] | 1,741,965,343,000 | 2025-03-14T15:18:16 | 307 | 0 | ---
{}
---
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
RAGPT-2_unfunctional - GGUF
- Model creator: https://huggingface.co/BueormLLC/
- Original model: https://huggingface.co/BueormLLC/RAGPT-2_unfunctional/
| Name | Quant method | Size |
| ---- | ---- | ---- |
| [RAGPT-2_unfunctional.Q2_K.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q2_K.gguf) | Q2_K | 0.08GB |
| [RAGPT-2_unfunctional.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.IQ3_XS.gguf) | IQ3_XS | 0.08GB |
| [RAGPT-2_unfunctional.IQ3_S.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.IQ3_S.gguf) | IQ3_S | 0.08GB |
| [RAGPT-2_unfunctional.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q3_K_S.gguf) | Q3_K_S | 0.08GB |
| [RAGPT-2_unfunctional.IQ3_M.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.IQ3_M.gguf) | IQ3_M | 0.09GB |
| [RAGPT-2_unfunctional.Q3_K.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q3_K.gguf) | Q3_K | 0.09GB |
| [RAGPT-2_unfunctional.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q3_K_M.gguf) | Q3_K_M | 0.09GB |
| [RAGPT-2_unfunctional.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q3_K_L.gguf) | Q3_K_L | 0.1GB |
| [RAGPT-2_unfunctional.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.IQ4_XS.gguf) | IQ4_XS | 0.1GB |
| [RAGPT-2_unfunctional.Q4_0.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q4_0.gguf) | Q4_0 | 0.1GB |
| [RAGPT-2_unfunctional.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.IQ4_NL.gguf) | IQ4_NL | 0.1GB |
| [RAGPT-2_unfunctional.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q4_K_S.gguf) | Q4_K_S | 0.1GB |
| [RAGPT-2_unfunctional.Q4_K.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q4_K.gguf) | Q4_K | 0.11GB |
| [RAGPT-2_unfunctional.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q4_K_M.gguf) | Q4_K_M | 0.11GB |
| [RAGPT-2_unfunctional.Q4_1.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q4_1.gguf) | Q4_1 | 0.11GB |
| [RAGPT-2_unfunctional.Q5_0.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q5_0.gguf) | Q5_0 | 0.11GB |
| [RAGPT-2_unfunctional.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q5_K_S.gguf) | Q5_K_S | 0.11GB |
| [RAGPT-2_unfunctional.Q5_K.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q5_K.gguf) | Q5_K | 0.12GB |
| [RAGPT-2_unfunctional.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q5_K_M.gguf) | Q5_K_M | 0.12GB |
| [RAGPT-2_unfunctional.Q5_1.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q5_1.gguf) | Q5_1 | 0.12GB |
| [RAGPT-2_unfunctional.Q6_K.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q6_K.gguf) | Q6_K | 0.13GB |
| [RAGPT-2_unfunctional.Q8_0.gguf](https://huggingface.co/RichardErkhov/BueormLLC_-_RAGPT-2_unfunctional-gguf/blob/main/RAGPT-2_unfunctional.Q8_0.gguf) | Q8_0 | 0.17GB |
Original model description:
---
license: mit
datasets:
- neural-bridge/rag-dataset-12000
- neural-bridge/rag-dataset-1200
language:
- en
---
# VERY IMPORTANT
- This model is in alpha phase and is NOT yet recommended for use.
- This model is obsolete today, recommended model [here.](https://huggingface.co/BueormLLC/RAGPT-2_Turbo)
# RAGPT-2 (unfunctional): Fine-tuned GPT-2 for Context-Based Question Answering
## Model Description
RAGPT-2 is a fine-tuned version of [GPT-2 small](https://huggingface.co/BueormLLC/CleanGPT), specifically adapted for context-based question answering tasks. This model has been trained to generate relevant answers based on a given context and question, similar to a Retrieval-Augmented Generation (RAG) system.
### Key Features
- Based on the GPT-2 small architecture (124M parameters)
- Fine-tuned on the "neural-bridge/rag-dataset-12000" and others dataset from Hugging Face
- Capable of generating answers based on provided context and questions
- Suitable for various question-answering applications
## Training Data
The model was fine-tuned using the "neural-bridge/rag-dataset-12000" and "neural-bridge/rag-dataset-1200" dataset, which contains:
- Context passages
- Questions related to the context
- Corresponding answers
## Fine-tuning Process
The fine-tuning process involved:
1. Loading the pre-trained GPT-2 small model
2. Preprocessing the dataset to combine context, question, and answer into a single text
3. Training the model to predict the next token given the context and question
### Hyperparameters
- Base model: GPT-2 small
- Number of training epochs: 8
- Batch size: 4
- Learning rate: Default AdamW optimizer settings
- Max sequence length: 512 tokens
## Usage
To use the model:
```python
from transformers import AutoTokenizer, AutoModelForCausalLM
tokenizer = AutoTokenizer.from_pretrained("BueormLLC/RAGPT-2_unfunctional")
model = AutoModelForCausalLM.from_pretrained("BueormLLC/RAGPT-2_unfunctional")
context = "Mount Everest is the highest mountain in the world, with a height of 8,848 meters."
question = "What is the height of Mount Everest?"
input_text = f"Context: {context}\nquestion: {question}\nanswer:"
input_ids = tokenizer.encode(input_text, return_tensors="pt")
output = model.generate(input_ids, max_length=150, num_return_sequences=1)
answer = tokenizer.decode(output[0], skip_special_tokens=True)
print(f"Respuesta generada: {answer}")
```
## Limitations
- The model's knowledge is limited to its training data and the base GPT-2 model.
- It may sometimes generate irrelevant or incorrect answers, especially for topics outside its training domain.
- The model does not have access to external information or real-time data.
## Ethical Considerations
Users should be aware that this model, like all language models, may reflect biases present in its training data. It should not be used as a sole source of information for critical decisions.
## Future Improvements
- Fine-tuning on a larger and more diverse dataset
- Experimenting with larger base models (e.g., GPT-2 medium or large)
- Implementing techniques to improve factual accuracy and reduce hallucinations
## Support us
- [Paypal](https://paypal.me/bueorm)
- [Patreon](https://patreon.com/bueorm)
### We appreciate your support, without you we could not do what we do.
## Citation
If you use this model in your research, please cite:
```
@misc{RAGPT,
author = {Bueorm},
title = {RAGPT-2: Fine-tuned GPT-2 for Context-Based Question Answering},
year = {2024},
publisher = {GitHub},
journal = {None},
howpublished = {\url{https://huggingface.co/BueormLLC/RAGPT-2_unfunctional}}
}
```
| [
"QUESTION_ANSWERING"
] | Non_BioNLP |
aasarmehdi/distilbert-base-uncased.finetuned-emotion | aasarmehdi | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,687,090,848,000 | 2023-06-18T15:12:34 | 8 | 0 | ---
datasets:
- emotion
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased.finetuned-emotion
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: emotion
type: emotion
config: split
split: validation
args: split
metrics:
- type: accuracy
value: 0.9285
name: Accuracy
- type: f1
value: 0.9285575296750973
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased.finetuned-emotion
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2139
- Accuracy: 0.9285
- F1: 0.9286
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.8378 | 1.0 | 250 | 0.3119 | 0.913 | 0.9104 |
| 0.2549 | 2.0 | 500 | 0.2139 | 0.9285 | 0.9286 |
### Framework versions
- Transformers 4.28.0
- Pytorch 1.12.1
- Datasets 2.12.0
- Tokenizers 0.11.0
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
Jiiiiiiiiiinw/finetuning-sentiment-model-3000-samples | Jiiiiiiiiiinw | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:imdb",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,681,376,197,000 | 2023-04-13T09:05:30 | 11 | 0 | ---
datasets:
- imdb
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: finetuning-sentiment-model-3000-samples
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: imdb
type: imdb
config: plain_text
split: test
args: plain_text
metrics:
- type: accuracy
value: 0.8866666666666667
name: Accuracy
- type: f1
value: 0.888157894736842
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# finetuning-sentiment-model-3000-samples
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3077
- Accuracy: 0.8867
- F1: 0.8882
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
### Framework versions
- Transformers 4.27.4
- Pytorch 2.0.0+cu118
- Datasets 2.11.0
- Tokenizers 0.13.3
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
Kevin123/distilbert-base-uncased-finetuned-cola | Kevin123 | text-classification | [
"transformers",
"pytorch",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:glue",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,663,880,639,000 | 2022-09-22T22:39:03 | 10 | 0 | ---
datasets:
- glue
license: apache-2.0
metrics:
- matthews_correlation
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-cola
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: glue
type: glue
args: cola
metrics:
- type: matthews_correlation
value: 0.5474713423103301
name: Matthews Correlation
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-cola
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the glue dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8663
- Matthews Correlation: 0.5475
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
### Training results
| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
|:-------------:|:-----:|:----:|:---------------:|:--------------------:|
| 0.5248 | 1.0 | 535 | 0.5171 | 0.4210 |
| 0.3418 | 2.0 | 1070 | 0.4971 | 0.5236 |
| 0.2289 | 3.0 | 1605 | 0.6874 | 0.5023 |
| 0.1722 | 4.0 | 2140 | 0.7680 | 0.5392 |
| 0.118 | 5.0 | 2675 | 0.8663 | 0.5475 |
### Framework versions
- Transformers 4.12.3
- Pytorch 1.8.1+cu102
- Datasets 1.18.3
- Tokenizers 0.10.3
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
nbogdan/flant5-small-1ex-paraphrasing-1epochs | nbogdan | null | [
"adapter-transformers",
"adapterhub:self-explanations",
"t5",
"dataset:self-explanations",
"region:us"
] | 1,693,842,308,000 | 2023-09-04T15:45:14 | 0 | 0 | ---
datasets:
- self-explanations
tags:
- adapterhub:self-explanations
- t5
- adapter-transformers
---
# Adapter `nbogdan/flant5-small-1ex-paraphrasing-1epochs` for google/flan-t5-small
An [adapter](https://adapterhub.ml) for the `google/flan-t5-small` model that was trained on the [self-explanations](https://adapterhub.ml/explore/self-explanations/) dataset.
This adapter was created for usage with the **[adapter-transformers](https://github.com/Adapter-Hub/adapter-transformers)** library.
## Usage
First, install `adapter-transformers`:
```
pip install -U adapter-transformers
```
_Note: adapter-transformers is a fork of transformers that acts as a drop-in replacement with adapter support. [More](https://docs.adapterhub.ml/installation.html)_
Now, the adapter can be loaded and activated like this:
```python
from transformers import AutoAdapterModel
model = AutoAdapterModel.from_pretrained("google/flan-t5-small")
adapter_name = model.load_adapter("nbogdan/flant5-small-1ex-paraphrasing-1epochs", source="hf", set_active=True)
```
## Architecture & Training
<!-- Add some description here -->
## Evaluation results
<!-- Add some description here -->
## Citation
<!-- Add some description here --> | [
"PARAPHRASING"
] | Non_BioNLP |
uboza10300/distilbert-hatexplain | uboza10300 | text-classification | [
"transformers",
"safetensors",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:hatexplain",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,733,805,785,000 | 2024-12-10T04:58:07 | 9 | 0 | ---
base_model: distilbert-base-uncased
datasets:
- hatexplain
library_name: transformers
license: apache-2.0
metrics:
- accuracy
- precision
- recall
- f1
tags:
- generated_from_trainer
model-index:
- name: distilbert-hatexplain
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: hatexplain
type: hatexplain
config: plain_text
split: validation
args: plain_text
metrics:
- type: accuracy
value: 0.6990644490644491
name: Accuracy
- type: precision
value: 0.6974890380019948
name: Precision
- type: recall
value: 0.6990644490644491
name: Recall
- type: f1
value: 0.6978790945993021
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-hatexplain
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the hatexplain dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8165
- Accuracy: 0.6991
- Precision: 0.6975
- Recall: 0.6991
- F1: 0.6979
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 0.6131 | 1.0 | 1923 | 0.7399 | 0.6925 | 0.6877 | 0.6925 | 0.6847 |
| 0.7386 | 2.0 | 3846 | 0.7254 | 0.7040 | 0.7033 | 0.7040 | 0.7036 |
| 0.6471 | 3.0 | 5769 | 0.8259 | 0.7019 | 0.6995 | 0.7019 | 0.7005 |
### Framework versions
- Transformers 4.47.0
- Pytorch 2.5.1+cu118
- Datasets 3.1.0
- Tokenizers 0.21.0
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
YakovElm/Qt15SetFitModel_balance_ratio_3 | YakovElm | text-classification | [
"sentence-transformers",
"pytorch",
"mpnet",
"setfit",
"text-classification",
"arxiv:2209.11055",
"license:apache-2.0",
"region:us"
] | 1,685,812,689,000 | 2023-06-03T17:18:44 | 10 | 0 | ---
license: apache-2.0
pipeline_tag: text-classification
tags:
- setfit
- sentence-transformers
- text-classification
---
# YakovElm/Qt15SetFitModel_balance_ratio_3
This is a [SetFit model](https://github.com/huggingface/setfit) that can be used for text classification. The model has been trained using an efficient few-shot learning technique that involves:
1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.
## Usage
To use this model for inference, first install the SetFit library:
```bash
python -m pip install setfit
```
You can then run inference as follows:
```python
from setfit import SetFitModel
# Download from Hub and run inference
model = SetFitModel.from_pretrained("YakovElm/Qt15SetFitModel_balance_ratio_3")
# Run inference
preds = model(["i loved the spiderman movie!", "pineapple on pizza is the worst 🤮"])
```
## BibTeX entry and citation info
```bibtex
@article{https://doi.org/10.48550/arxiv.2209.11055,
doi = {10.48550/ARXIV.2209.11055},
url = {https://arxiv.org/abs/2209.11055},
author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
title = {Efficient Few-Shot Learning Without Prompts},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}
}
```
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
farleyknight/patent-summarization-google-bigbird-pegasus-large-arxiv-2022-09-20 | farleyknight | text2text-generation | [
"transformers",
"pytorch",
"bigbird_pegasus",
"text2text-generation",
"generated_from_trainer",
"dataset:farleyknight/big_patent_5_percent",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,663,709,552,000 | 2022-09-23T02:53:23 | 40 | 0 | ---
datasets:
- farleyknight/big_patent_5_percent
license: apache-2.0
metrics:
- rouge
tags:
- generated_from_trainer
model-index:
- name: patent-summarization-google-bigbird-pegasus-large-arxiv-2022-09-20
results:
- task:
type: summarization
name: Summarization
dataset:
name: farleyknight/big_patent_5_percent
type: farleyknight/big_patent_5_percent
config: all
split: train
args: all
metrics:
- type: rouge
value: 37.3764
name: Rouge1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# patent-summarization-google-bigbird-pegasus-large-arxiv-2022-09-20
This model is a fine-tuned version of [google/bigbird-pegasus-large-arxiv](https://huggingface.co/google/bigbird-pegasus-large-arxiv) on the farleyknight/big_patent_5_percent dataset.
It achieves the following results on the evaluation set:
- Loss: 2.2617
- Rouge1: 37.3764
- Rouge2: 13.2442
- Rougel: 26.011
- Rougelsum: 31.0145
- Gen Len: 113.8789
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 1
- eval_batch_size: 1
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 1.0
### Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|:-------------:|:-----:|:-----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:|
| 2.6121 | 0.08 | 5000 | 2.5652 | 35.0673 | 12.0073 | 24.5471 | 28.9315 | 119.9866 |
| 2.5182 | 0.17 | 10000 | 2.4797 | 34.6909 | 11.6432 | 24.87 | 28.1543 | 119.2043 |
| 2.5102 | 0.25 | 15000 | 2.4238 | 35.8574 | 12.2402 | 25.0712 | 29.5607 | 115.2890 |
| 2.4292 | 0.33 | 20000 | 2.3869 | 36.0133 | 12.2453 | 25.4039 | 29.483 | 112.5920 |
| 2.3678 | 0.41 | 25000 | 2.3594 | 35.238 | 11.6833 | 25.0449 | 28.3313 | 119.1739 |
| 2.3511 | 0.5 | 30000 | 2.3326 | 36.7755 | 12.8394 | 25.7218 | 30.2594 | 110.5819 |
| 2.3334 | 0.58 | 35000 | 2.3125 | 36.6317 | 12.7493 | 25.5388 | 30.094 | 115.5998 |
| 2.3833 | 0.66 | 40000 | 2.2943 | 37.1219 | 13.1564 | 25.7571 | 30.8666 | 113.8222 |
| 2.341 | 0.75 | 45000 | 2.2813 | 36.4962 | 12.6225 | 25.6904 | 29.9741 | 115.9845 |
| 2.3179 | 0.83 | 50000 | 2.2725 | 37.3535 | 13.1596 | 25.7385 | 31.056 | 117.7754 |
| 2.3164 | 0.91 | 55000 | 2.2654 | 36.9191 | 12.9316 | 25.7586 | 30.4691 | 116.1670 |
| 2.3046 | 0.99 | 60000 | 2.2618 | 37.3992 | 13.2731 | 26.0327 | 31.0338 | 114.5195 |
### Framework versions
- Transformers 4.23.0.dev0
- Pytorch 1.12.0
- Datasets 2.4.0
- Tokenizers 0.12.1
| [
"SUMMARIZATION"
] | Non_BioNLP |
poltextlab/xlm-roberta-large-english-social-cap-v3 | poltextlab | null | [
"pytorch",
"xlm-roberta",
"arxiv:1910.09700",
"region:us"
] | 1,729,244,656,000 | 2025-02-26T16:08:16 | 99 | 0 | ---
extra_gated_prompt: 'Our models are intended for academic use only. If you are not
affiliated with an academic institution, please provide a rationale for using our
models. Please allow us a few business days to manually review subscriptions.
If you use our models for your work or research, please cite this paper: Sebők,
M., Máté, Á., Ring, O., Kovács, V., & Lehoczki, R. (2024). Leveraging Open Large
Language Models for Multilingual Policy Topic Classification: The Babel Machine
Approach. Social Science Computer Review, 0(0). https://doi.org/10.1177/08944393241259434'
extra_gated_fields:
Name: text
Country: country
Institution: text
Institution Email: text
Please specify your academic use case: text
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
## Model Details
The translation table from the model results to CAP codes is the following:
```python
CAP_NUM_DICT = {
0: 1,
1: 2,
2: 3,
3: 4,
4: 5,
5: 6,
6: 7,
7: 8,
8: 9,
9: 10,
10: 12,
11: 13,
12: 14,
13: 15,
14: 16,
15: 17,
16: 18,
17: 19,
18: 20,
19: 21,
20: 23,
21: 999,
}
```
### Model Description
<!-- Provide a longer summary of what this model is. -->
- **Developed by:** [More Information Needed]
- **Funded by [optional]:** [More Information Needed]
- **Shared by [optional]:** [More Information Needed]
- **Model type:** [More Information Needed]
- **Language(s) (NLP):** [More Information Needed]
- **License:** [More Information Needed]
- **Finetuned from model [optional]:** [More Information Needed]
### Model Sources [optional]
<!-- Provide the basic links for the model. -->
- **Repository:** [More Information Needed]
- **Paper [optional]:** [More Information Needed]
- **Demo [optional]:** [More Information Needed]
## Uses
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
### Direct Use
<!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
[More Information Needed]
### Downstream Use [optional]
<!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
[More Information Needed]
### Out-of-Scope Use
<!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
[More Information Needed]
## Bias, Risks, and Limitations
<!-- This section is meant to convey both technical and sociotechnical limitations. -->
[More Information Needed]
### Recommendations
<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
## How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
## Training Details
### Training Data
<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
[More Information Needed]
### Training Procedure
<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
#### Preprocessing [optional]
[More Information Needed]
#### Training Hyperparameters
- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
#### Speeds, Sizes, Times [optional]
<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
[More Information Needed]
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
### Testing Data, Factors & Metrics
#### Testing Data
<!-- This should link to a Dataset Card if possible. -->
[More Information Needed]
#### Factors
<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
[More Information Needed]
#### Metrics
<!-- These are the evaluation metrics being used, ideally with a description of why. -->
[More Information Needed]
### Results
[More Information Needed]
#### Summary
## Model Examination [optional]
<!-- Relevant interpretability work for the model goes here -->
[More Information Needed]
## Environmental Impact
<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]
## Technical Specifications [optional]
### Model Architecture and Objective
[More Information Needed]
### Compute Infrastructure
[More Information Needed]
#### Hardware
[More Information Needed]
#### Software
[More Information Needed]
## Citation [optional]
<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
**BibTeX:**
[More Information Needed]
**APA:**
[More Information Needed]
## Glossary [optional]
<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
[More Information Needed]
## More Information [optional]
[More Information Needed]
## Model Card Authors [optional]
[More Information Needed]
## Model Card Contact
[More Information Needed] | [
"TRANSLATION"
] | Non_BioNLP |
TripleH/distilbert-base-uncased-finetuned-emotion | TripleH | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,675,439,119,000 | 2023-02-03T16:26:23 | 114 | 0 | ---
datasets:
- emotion
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-emotion
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: emotion
type: emotion
config: split
split: validation
args: split
metrics:
- type: accuracy
value: 0.925
name: Accuracy
- type: f1
value: 0.9250927884813909
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-emotion
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2157
- Accuracy: 0.925
- F1: 0.9251
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.8015 | 1.0 | 250 | 0.3087 | 0.902 | 0.8989 |
| 0.2414 | 2.0 | 500 | 0.2157 | 0.925 | 0.9251 |
### Framework versions
- Transformers 4.26.0
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
haophancs/bge-m3-financial-matryoshka | haophancs | sentence-similarity | [
"sentence-transformers",
"onnx",
"safetensors",
"xlm-roberta",
"sentence-similarity",
"feature-extraction",
"generated_from_trainer",
"dataset_size:6300",
"loss:MatryoshkaLoss",
"loss:MultipleNegativesRankingLoss",
"en",
"arxiv:1908.10084",
"arxiv:2205.13147",
"arxiv:1705.00652",
"base_model:BAAI/bge-m3",
"base_model:quantized:BAAI/bge-m3",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
] | 1,719,057,886,000 | 2024-07-09T04:46:45 | 37 | 1 | ---
base_model: BAAI/bge-m3
datasets: []
language:
- en
library_name: sentence-transformers
license: apache-2.0
metrics:
- cosine_accuracy@1
- cosine_accuracy@3
- cosine_accuracy@5
- cosine_accuracy@10
- cosine_precision@1
- cosine_precision@3
- cosine_precision@5
- cosine_precision@10
- cosine_recall@1
- cosine_recall@3
- cosine_recall@5
- cosine_recall@10
- cosine_ndcg@10
- cosine_mrr@10
- cosine_map@100
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- generated_from_trainer
- dataset_size:6300
- loss:MatryoshkaLoss
- loss:MultipleNegativesRankingLoss
widget:
- source_sentence: The consolidated financial statements and accompanying notes listed
in Part IV, Item 15(a)(1) of this Annual Report on Form 10-K.
sentences:
- How much total space does an average The Home Depot store encompass including
its garden area?
- What section of the Annual Report on Form 10-K contains the consolidated financial
statements and accompanying notes?
- What types of competitive factors does Garmin believe are important in its markets?
- source_sentence: Item 3. Legal Proceedings, which covers litigation and regulatory
matters, refers to Note 12 – Commitments and Contingencies for more detailed information
within the Consolidated Financial Statements.
sentences:
- What pages contain the Financial Statements and Supplementary Data in IBM’s 2023
Annual Report to Stockholders?
- In which note can further details on Legal Proceedings be found within the Consolidated
Financial Statements?
- What is the title of Item 8 in the document?
- source_sentence: Net Revenues for the Entertainment segment were $659.3 million
in 2023.
sentences:
- What were the net revenues for the Entertainment segment in 2023?
- How much net cash was provided by operating activities in 2023?
- What was the net income reported for the fiscal year ending in August 2023?
- source_sentence: 'The capital allocation program focuses on three objectives: (1)
grow our business at an average target ROIC-adjusted rate of 20% or greater; (2)
maintain a strong investment-grade balance sheet, including a target average automotive
cash balance of $18.0 billion; and (3) after the first two objectives are met,
return available cash to shareholders.'
sentences:
- Why is ICE Mortgage Technology subject to the examination by the Federal Financial
Institutions Examination Council (FFIEC) and its member agencies?
- What type of regulations do U.S. automobiles need to comply with under the National
Highway Traffic Safety Administration?
- What are the three objectives of the capital allocation program referenced?
- source_sentence: As of January 28, 2024 the net carrying value of our inventories
was $1.3 billion, which included provisions for obsolete and damaged inventory
of $139.7 million.
sentences:
- What is the status of the company's inventory as of January 28, 2024, in terms
of its valuation and provisions for obsolescence?
- What is the relationship between the ESG goals and the long-term growth strategy?
- What were the financial impacts of Ford's investments in Rivian and Argo in the
year 2022?
model-index:
- name: BGE-M3 Financial Matryoshka
results:
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 1024
type: dim_1024
metrics:
- type: cosine_accuracy@1
value: 0.7171428571428572
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.8314285714285714
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.87
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.9142857142857143
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.7171428571428572
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.27714285714285714
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.174
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.09142857142857141
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.7171428571428572
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.8314285714285714
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.87
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.9142857142857143
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.8152097277196483
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.7835873015873015
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.7867088346410263
name: Cosine Map@100
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 768
type: dim_768
metrics:
- type: cosine_accuracy@1
value: 0.7128571428571429
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.8342857142857143
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.8657142857142858
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.91
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.7128571428571429
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.2780952380952381
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.17314285714285713
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.09099999999999998
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.7128571428571429
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.8342857142857143
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.8657142857142858
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.91
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.8122143155463835
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.7808730158730155
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.7843065190190194
name: Cosine Map@100
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 512
type: dim_512
metrics:
- type: cosine_accuracy@1
value: 0.7114285714285714
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.8357142857142857
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.8642857142857143
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.91
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.7114285714285714
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.2785714285714286
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.17285714285714285
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.09099999999999998
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.7114285714285714
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.8357142857142857
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.8642857142857143
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.91
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.8109635546819154
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.7792959183673466
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.782703758965192
name: Cosine Map@100
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: dim 384
type: dim_384
metrics:
- type: cosine_accuracy@1
value: 0.7142857142857143
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.8328571428571429
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.8628571428571429
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.9128571428571428
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.7142857142857143
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.2776190476190476
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.17257142857142854
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.09128571428571428
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.7142857142857143
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.8328571428571429
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.8628571428571429
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.9128571428571428
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.8125530857386527
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.7806292517006799
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.7837508100457361
name: Cosine Map@100
---
# BGE-M3 Financial Matryoshka
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3). It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
## Model Details
### Model Description
- **Model Type:** Sentence Transformer
- **Base model:** [BAAI/bge-m3](https://huggingface.co/BAAI/bge-m3) <!-- at revision babcf60cae0a1f438d7ade582983d4ba462303c2 -->
- **Maximum Sequence Length:** 8192 tokens
- **Output Dimensionality:** 1024 tokens
- **Similarity Function:** Cosine Similarity
<!-- - **Training Dataset:** Unknown -->
- **Language:** en
- **License:** apache-2.0
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 8192, 'do_lower_case': False}) with Transformer model: XLMRobertaModel
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("haophancs/bge-m3-financial-matryoshka")
# Run inference
sentences = [
'As of January 28, 2024 the net carrying value of our inventories was $1.3 billion, which included provisions for obsolete and damaged inventory of $139.7 million.',
"What is the status of the company's inventory as of January 28, 2024, in terms of its valuation and provisions for obsolescence?",
'What is the relationship between the ESG goals and the long-term growth strategy?',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
```
<!--
### Direct Usage (Transformers)
<details><summary>Click to see the direct usage in Transformers</summary>
</details>
-->
<!--
### Downstream Usage (Sentence Transformers)
You can finetune this model on your own dataset.
<details><summary>Click to expand</summary>
</details>
-->
<!--
### Out-of-Scope Use
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->
## Evaluation
### Metrics
#### Information Retrieval
* Dataset: `dim_1024`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.7171 |
| cosine_accuracy@3 | 0.8314 |
| cosine_accuracy@5 | 0.87 |
| cosine_accuracy@10 | 0.9143 |
| cosine_precision@1 | 0.7171 |
| cosine_precision@3 | 0.2771 |
| cosine_precision@5 | 0.174 |
| cosine_precision@10 | 0.0914 |
| cosine_recall@1 | 0.7171 |
| cosine_recall@3 | 0.8314 |
| cosine_recall@5 | 0.87 |
| cosine_recall@10 | 0.9143 |
| cosine_ndcg@10 | 0.8152 |
| cosine_mrr@10 | 0.7836 |
| **cosine_map@100** | **0.7867** |
#### Information Retrieval
* Dataset: `dim_768`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.7129 |
| cosine_accuracy@3 | 0.8343 |
| cosine_accuracy@5 | 0.8657 |
| cosine_accuracy@10 | 0.91 |
| cosine_precision@1 | 0.7129 |
| cosine_precision@3 | 0.2781 |
| cosine_precision@5 | 0.1731 |
| cosine_precision@10 | 0.091 |
| cosine_recall@1 | 0.7129 |
| cosine_recall@3 | 0.8343 |
| cosine_recall@5 | 0.8657 |
| cosine_recall@10 | 0.91 |
| cosine_ndcg@10 | 0.8122 |
| cosine_mrr@10 | 0.7809 |
| **cosine_map@100** | **0.7843** |
#### Information Retrieval
* Dataset: `dim_512`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.7114 |
| cosine_accuracy@3 | 0.8357 |
| cosine_accuracy@5 | 0.8643 |
| cosine_accuracy@10 | 0.91 |
| cosine_precision@1 | 0.7114 |
| cosine_precision@3 | 0.2786 |
| cosine_precision@5 | 0.1729 |
| cosine_precision@10 | 0.091 |
| cosine_recall@1 | 0.7114 |
| cosine_recall@3 | 0.8357 |
| cosine_recall@5 | 0.8643 |
| cosine_recall@10 | 0.91 |
| cosine_ndcg@10 | 0.811 |
| cosine_mrr@10 | 0.7793 |
| **cosine_map@100** | **0.7827** |
#### Information Retrieval
* Dataset: `dim_384`
* Evaluated with [<code>InformationRetrievalEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.7143 |
| cosine_accuracy@3 | 0.8329 |
| cosine_accuracy@5 | 0.8629 |
| cosine_accuracy@10 | 0.9129 |
| cosine_precision@1 | 0.7143 |
| cosine_precision@3 | 0.2776 |
| cosine_precision@5 | 0.1726 |
| cosine_precision@10 | 0.0913 |
| cosine_recall@1 | 0.7143 |
| cosine_recall@3 | 0.8329 |
| cosine_recall@5 | 0.8629 |
| cosine_recall@10 | 0.9129 |
| cosine_ndcg@10 | 0.8126 |
| cosine_mrr@10 | 0.7806 |
| **cosine_map@100** | **0.7838** |
<!--
## Bias, Risks and Limitations
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->
<!--
### Recommendations
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->
## Training Details
### Training Dataset
#### Unnamed Dataset
* Size: 6,300 training samples
* Columns: <code>positive</code> and <code>anchor</code>
* Approximate statistics based on the first 1000 samples:
| | positive | anchor |
|:--------|:-------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|
| type | string | string |
| details | <ul><li>min: 11 tokens</li><li>mean: 51.97 tokens</li><li>max: 1146 tokens</li></ul> | <ul><li>min: 7 tokens</li><li>mean: 21.63 tokens</li><li>max: 47 tokens</li></ul> |
* Samples:
| positive | anchor |
|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------------------------------------------|
| <code>From fiscal year 2022 to 2023, the cost of revenue as a percentage of total net revenue decreased by 3 percent.</code> | <code>What was the percentage change in cost of revenue as a percentage of total net revenue from fiscal year 2022 to 2023?</code> |
| <code> •Operating income increased $321 million, or 2%, to $18.1 billion versus year ago due to the increase in net sales, partially offset by a modest decrease in operating margin.</code> | <code>What factors contributed to the increase in operating income for Procter & Gamble in 2023?</code> |
| <code>market specific brands including 'Aurrera,' 'Lider,' and 'PhonePe.'</code> | <code>What specific brands does Walmart International market?</code> |
* Loss: [<code>MatryoshkaLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#matryoshkaloss) with these parameters:
```json
{
"loss": "MultipleNegativesRankingLoss",
"matryoshka_dims": [
1024,
768,
512,
384
],
"matryoshka_weights": [
1,
1,
1,
1
],
"n_dims_per_step": -1
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: epoch
- `per_device_train_batch_size`: 4
- `per_device_eval_batch_size`: 2
- `gradient_accumulation_steps`: 2
- `learning_rate`: 2e-05
- `num_train_epochs`: 4
- `lr_scheduler_type`: cosine
- `warmup_ratio`: 0.1
- `bf16`: True
- `tf32`: True
- `load_best_model_at_end`: True
- `optim`: adamw_torch_fused
- `batch_sampler`: no_duplicates
#### All Hyperparameters
<details><summary>Click to expand</summary>
- `overwrite_output_dir`: False
- `do_predict`: False
- `eval_strategy`: epoch
- `prediction_loss_only`: True
- `per_device_train_batch_size`: 4
- `per_device_eval_batch_size`: 2
- `per_gpu_train_batch_size`: None
- `per_gpu_eval_batch_size`: None
- `gradient_accumulation_steps`: 2
- `eval_accumulation_steps`: None
- `learning_rate`: 2e-05
- `weight_decay`: 0.0
- `adam_beta1`: 0.9
- `adam_beta2`: 0.999
- `adam_epsilon`: 1e-08
- `max_grad_norm`: 1.0
- `num_train_epochs`: 4
- `max_steps`: -1
- `lr_scheduler_type`: cosine
- `lr_scheduler_kwargs`: {}
- `warmup_ratio`: 0.1
- `warmup_steps`: 0
- `log_level`: passive
- `log_level_replica`: warning
- `log_on_each_node`: True
- `logging_nan_inf_filter`: True
- `save_safetensors`: True
- `save_on_each_node`: False
- `save_only_model`: False
- `restore_callback_states_from_checkpoint`: False
- `no_cuda`: False
- `use_cpu`: False
- `use_mps_device`: False
- `seed`: 42
- `data_seed`: None
- `jit_mode_eval`: False
- `use_ipex`: False
- `bf16`: True
- `fp16`: False
- `fp16_opt_level`: O1
- `half_precision_backend`: auto
- `bf16_full_eval`: False
- `fp16_full_eval`: False
- `tf32`: True
- `local_rank`: 0
- `ddp_backend`: None
- `tpu_num_cores`: None
- `tpu_metrics_debug`: False
- `debug`: []
- `dataloader_drop_last`: False
- `dataloader_num_workers`: 0
- `dataloader_prefetch_factor`: None
- `past_index`: -1
- `disable_tqdm`: False
- `remove_unused_columns`: True
- `label_names`: None
- `load_best_model_at_end`: True
- `ignore_data_skip`: False
- `fsdp`: []
- `fsdp_min_num_params`: 0
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
- `fsdp_transformer_layer_cls_to_wrap`: None
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
- `deepspeed`: None
- `label_smoothing_factor`: 0.0
- `optim`: adamw_torch_fused
- `optim_args`: None
- `adafactor`: False
- `group_by_length`: False
- `length_column_name`: length
- `ddp_find_unused_parameters`: None
- `ddp_bucket_cap_mb`: None
- `ddp_broadcast_buffers`: False
- `dataloader_pin_memory`: True
- `dataloader_persistent_workers`: False
- `skip_memory_metrics`: True
- `use_legacy_prediction_loop`: False
- `push_to_hub`: False
- `resume_from_checkpoint`: None
- `hub_model_id`: None
- `hub_strategy`: every_save
- `hub_private_repo`: False
- `hub_always_push`: False
- `gradient_checkpointing`: False
- `gradient_checkpointing_kwargs`: None
- `include_inputs_for_metrics`: False
- `eval_do_concat_batches`: True
- `fp16_backend`: auto
- `push_to_hub_model_id`: None
- `push_to_hub_organization`: None
- `mp_parameters`:
- `auto_find_batch_size`: False
- `full_determinism`: False
- `torchdynamo`: None
- `ray_scope`: last
- `ddp_timeout`: 1800
- `torch_compile`: False
- `torch_compile_backend`: None
- `torch_compile_mode`: None
- `dispatch_batches`: None
- `split_batches`: None
- `include_tokens_per_second`: False
- `include_num_input_tokens_seen`: False
- `neftune_noise_alpha`: None
- `optim_target_modules`: None
- `batch_eval_metrics`: False
- `batch_sampler`: no_duplicates
- `multi_dataset_batch_sampler`: proportional
</details>
### Training Logs
<details><summary>Click to expand</summary>
| Epoch | Step | Training Loss | dim_1024_cosine_map@100 | dim_384_cosine_map@100 | dim_512_cosine_map@100 | dim_768_cosine_map@100 |
|:----------:|:--------:|:-------------:|:-----------------------:|:----------------------:|:----------------------:|:----------------------:|
| 0.0127 | 10 | 0.2059 | - | - | - | - |
| 0.0254 | 20 | 0.2612 | - | - | - | - |
| 0.0381 | 30 | 0.0873 | - | - | - | - |
| 0.0508 | 40 | 0.1352 | - | - | - | - |
| 0.0635 | 50 | 0.156 | - | - | - | - |
| 0.0762 | 60 | 0.0407 | - | - | - | - |
| 0.0889 | 70 | 0.09 | - | - | - | - |
| 0.1016 | 80 | 0.027 | - | - | - | - |
| 0.1143 | 90 | 0.0978 | - | - | - | - |
| 0.1270 | 100 | 0.0105 | - | - | - | - |
| 0.1397 | 110 | 0.0402 | - | - | - | - |
| 0.1524 | 120 | 0.0745 | - | - | - | - |
| 0.1651 | 130 | 0.0655 | - | - | - | - |
| 0.1778 | 140 | 0.0075 | - | - | - | - |
| 0.1905 | 150 | 0.0141 | - | - | - | - |
| 0.2032 | 160 | 0.0615 | - | - | - | - |
| 0.2159 | 170 | 0.0029 | - | - | - | - |
| 0.2286 | 180 | 0.0269 | - | - | - | - |
| 0.2413 | 190 | 0.0724 | - | - | - | - |
| 0.2540 | 200 | 0.0218 | - | - | - | - |
| 0.2667 | 210 | 0.0027 | - | - | - | - |
| 0.2794 | 220 | 0.007 | - | - | - | - |
| 0.2921 | 230 | 0.0814 | - | - | - | - |
| 0.3048 | 240 | 0.0326 | - | - | - | - |
| 0.3175 | 250 | 0.0061 | - | - | - | - |
| 0.3302 | 260 | 0.0471 | - | - | - | - |
| 0.3429 | 270 | 0.0115 | - | - | - | - |
| 0.3556 | 280 | 0.0021 | - | - | - | - |
| 0.3683 | 290 | 0.0975 | - | - | - | - |
| 0.3810 | 300 | 0.0572 | - | - | - | - |
| 0.3937 | 310 | 0.0125 | - | - | - | - |
| 0.4063 | 320 | 0.04 | - | - | - | - |
| 0.4190 | 330 | 0.0023 | - | - | - | - |
| 0.4317 | 340 | 0.0121 | - | - | - | - |
| 0.4444 | 350 | 0.0116 | - | - | - | - |
| 0.4571 | 360 | 0.0059 | - | - | - | - |
| 0.4698 | 370 | 0.0217 | - | - | - | - |
| 0.4825 | 380 | 0.0294 | - | - | - | - |
| 0.4952 | 390 | 0.1102 | - | - | - | - |
| 0.5079 | 400 | 0.0103 | - | - | - | - |
| 0.5206 | 410 | 0.0023 | - | - | - | - |
| 0.5333 | 420 | 0.0157 | - | - | - | - |
| 0.5460 | 430 | 0.0805 | - | - | - | - |
| 0.5587 | 440 | 0.0168 | - | - | - | - |
| 0.5714 | 450 | 0.1279 | - | - | - | - |
| 0.5841 | 460 | 0.2012 | - | - | - | - |
| 0.5968 | 470 | 0.0436 | - | - | - | - |
| 0.6095 | 480 | 0.0204 | - | - | - | - |
| 0.6222 | 490 | 0.0097 | - | - | - | - |
| 0.6349 | 500 | 0.0013 | - | - | - | - |
| 0.6476 | 510 | 0.0042 | - | - | - | - |
| 0.6603 | 520 | 0.0034 | - | - | - | - |
| 0.6730 | 530 | 0.0226 | - | - | - | - |
| 0.6857 | 540 | 0.0267 | - | - | - | - |
| 0.6984 | 550 | 0.0007 | - | - | - | - |
| 0.7111 | 560 | 0.0766 | - | - | - | - |
| 0.7238 | 570 | 0.2174 | - | - | - | - |
| 0.7365 | 580 | 0.0089 | - | - | - | - |
| 0.7492 | 590 | 0.0794 | - | - | - | - |
| 0.7619 | 600 | 0.0031 | - | - | - | - |
| 0.7746 | 610 | 0.0499 | - | - | - | - |
| 0.7873 | 620 | 0.0105 | - | - | - | - |
| 0.8 | 630 | 0.0097 | - | - | - | - |
| 0.8127 | 640 | 0.0028 | - | - | - | - |
| 0.8254 | 650 | 0.0029 | - | - | - | - |
| 0.8381 | 660 | 0.1811 | - | - | - | - |
| 0.8508 | 670 | 0.064 | - | - | - | - |
| 0.8635 | 680 | 0.0139 | - | - | - | - |
| 0.8762 | 690 | 0.055 | - | - | - | - |
| 0.8889 | 700 | 0.0013 | - | - | - | - |
| 0.9016 | 710 | 0.0402 | - | - | - | - |
| 0.9143 | 720 | 0.0824 | - | - | - | - |
| 0.9270 | 730 | 0.03 | - | - | - | - |
| 0.9397 | 740 | 0.0337 | - | - | - | - |
| 0.9524 | 750 | 0.1192 | - | - | - | - |
| 0.9651 | 760 | 0.0039 | - | - | - | - |
| 0.9778 | 770 | 0.004 | - | - | - | - |
| 0.9905 | 780 | 0.1413 | - | - | - | - |
| 0.9994 | 787 | - | 0.7851 | 0.7794 | 0.7822 | 0.7863 |
| 1.0032 | 790 | 0.019 | - | - | - | - |
| 1.0159 | 800 | 0.0587 | - | - | - | - |
| 1.0286 | 810 | 0.0186 | - | - | - | - |
| 1.0413 | 820 | 0.0018 | - | - | - | - |
| 1.0540 | 830 | 0.0631 | - | - | - | - |
| 1.0667 | 840 | 0.0127 | - | - | - | - |
| 1.0794 | 850 | 0.0037 | - | - | - | - |
| 1.0921 | 860 | 0.0029 | - | - | - | - |
| 1.1048 | 870 | 0.1437 | - | - | - | - |
| 1.1175 | 880 | 0.0015 | - | - | - | - |
| 1.1302 | 890 | 0.0024 | - | - | - | - |
| 1.1429 | 900 | 0.0133 | - | - | - | - |
| 1.1556 | 910 | 0.0245 | - | - | - | - |
| 1.1683 | 920 | 0.0017 | - | - | - | - |
| 1.1810 | 930 | 0.0007 | - | - | - | - |
| 1.1937 | 940 | 0.002 | - | - | - | - |
| 1.2063 | 950 | 0.0044 | - | - | - | - |
| 1.2190 | 960 | 0.0009 | - | - | - | - |
| 1.2317 | 970 | 0.01 | - | - | - | - |
| 1.2444 | 980 | 0.0026 | - | - | - | - |
| 1.2571 | 990 | 0.0017 | - | - | - | - |
| 1.2698 | 1000 | 0.0014 | - | - | - | - |
| 1.2825 | 1010 | 0.0009 | - | - | - | - |
| 1.2952 | 1020 | 0.0829 | - | - | - | - |
| 1.3079 | 1030 | 0.0011 | - | - | - | - |
| 1.3206 | 1040 | 0.012 | - | - | - | - |
| 1.3333 | 1050 | 0.0019 | - | - | - | - |
| 1.3460 | 1060 | 0.0007 | - | - | - | - |
| 1.3587 | 1070 | 0.0141 | - | - | - | - |
| 1.3714 | 1080 | 0.0003 | - | - | - | - |
| 1.3841 | 1090 | 0.001 | - | - | - | - |
| 1.3968 | 1100 | 0.0005 | - | - | - | - |
| 1.4095 | 1110 | 0.0031 | - | - | - | - |
| 1.4222 | 1120 | 0.0004 | - | - | - | - |
| 1.4349 | 1130 | 0.0054 | - | - | - | - |
| 1.4476 | 1140 | 0.0003 | - | - | - | - |
| 1.4603 | 1150 | 0.0007 | - | - | - | - |
| 1.4730 | 1160 | 0.0009 | - | - | - | - |
| 1.4857 | 1170 | 0.001 | - | - | - | - |
| 1.4984 | 1180 | 0.0006 | - | - | - | - |
| 1.5111 | 1190 | 0.0046 | - | - | - | - |
| 1.5238 | 1200 | 0.0003 | - | - | - | - |
| 1.5365 | 1210 | 0.0002 | - | - | - | - |
| 1.5492 | 1220 | 0.004 | - | - | - | - |
| 1.5619 | 1230 | 0.0017 | - | - | - | - |
| 1.5746 | 1240 | 0.0003 | - | - | - | - |
| 1.5873 | 1250 | 0.0027 | - | - | - | - |
| 1.6 | 1260 | 0.1134 | - | - | - | - |
| 1.6127 | 1270 | 0.0007 | - | - | - | - |
| 1.6254 | 1280 | 0.0005 | - | - | - | - |
| 1.6381 | 1290 | 0.0008 | - | - | - | - |
| 1.6508 | 1300 | 0.0001 | - | - | - | - |
| 1.6635 | 1310 | 0.0023 | - | - | - | - |
| 1.6762 | 1320 | 0.0005 | - | - | - | - |
| 1.6889 | 1330 | 0.0004 | - | - | - | - |
| 1.7016 | 1340 | 0.0003 | - | - | - | - |
| 1.7143 | 1350 | 0.0347 | - | - | - | - |
| 1.7270 | 1360 | 0.0339 | - | - | - | - |
| 1.7397 | 1370 | 0.0003 | - | - | - | - |
| 1.7524 | 1380 | 0.0005 | - | - | - | - |
| 1.7651 | 1390 | 0.0002 | - | - | - | - |
| 1.7778 | 1400 | 0.0031 | - | - | - | - |
| 1.7905 | 1410 | 0.0002 | - | - | - | - |
| 1.8032 | 1420 | 0.0012 | - | - | - | - |
| 1.8159 | 1430 | 0.0002 | - | - | - | - |
| 1.8286 | 1440 | 0.0002 | - | - | - | - |
| 1.8413 | 1450 | 0.0004 | - | - | - | - |
| 1.8540 | 1460 | 0.011 | - | - | - | - |
| 1.8667 | 1470 | 0.0824 | - | - | - | - |
| 1.8794 | 1480 | 0.0003 | - | - | - | - |
| 1.8921 | 1490 | 0.0004 | - | - | - | - |
| 1.9048 | 1500 | 0.0006 | - | - | - | - |
| 1.9175 | 1510 | 0.015 | - | - | - | - |
| 1.9302 | 1520 | 0.0004 | - | - | - | - |
| 1.9429 | 1530 | 0.0004 | - | - | - | - |
| 1.9556 | 1540 | 0.0011 | - | - | - | - |
| 1.9683 | 1550 | 0.0003 | - | - | - | - |
| 1.9810 | 1560 | 0.0006 | - | - | - | - |
| 1.9937 | 1570 | 0.0042 | - | - | - | - |
| 2.0 | 1575 | - | 0.7862 | 0.7855 | 0.7852 | 0.7878 |
| 2.0063 | 1580 | 0.0005 | - | - | - | - |
| 2.0190 | 1590 | 0.002 | - | - | - | - |
| 2.0317 | 1600 | 0.0013 | - | - | - | - |
| 2.0444 | 1610 | 0.0002 | - | - | - | - |
| 2.0571 | 1620 | 0.0035 | - | - | - | - |
| 2.0698 | 1630 | 0.0004 | - | - | - | - |
| 2.0825 | 1640 | 0.0002 | - | - | - | - |
| 2.0952 | 1650 | 0.0032 | - | - | - | - |
| 2.1079 | 1660 | 0.0916 | - | - | - | - |
| 2.1206 | 1670 | 0.0002 | - | - | - | - |
| 2.1333 | 1680 | 0.0006 | - | - | - | - |
| 2.1460 | 1690 | 0.0002 | - | - | - | - |
| 2.1587 | 1700 | 0.0003 | - | - | - | - |
| 2.1714 | 1710 | 0.0001 | - | - | - | - |
| 2.1841 | 1720 | 0.0001 | - | - | - | - |
| 2.1968 | 1730 | 0.0004 | - | - | - | - |
| 2.2095 | 1740 | 0.0004 | - | - | - | - |
| 2.2222 | 1750 | 0.0001 | - | - | - | - |
| 2.2349 | 1760 | 0.0002 | - | - | - | - |
| 2.2476 | 1770 | 0.0007 | - | - | - | - |
| 2.2603 | 1780 | 0.0001 | - | - | - | - |
| 2.2730 | 1790 | 0.0002 | - | - | - | - |
| 2.2857 | 1800 | 0.0004 | - | - | - | - |
| 2.2984 | 1810 | 0.0711 | - | - | - | - |
| 2.3111 | 1820 | 0.0001 | - | - | - | - |
| 2.3238 | 1830 | 0.0005 | - | - | - | - |
| 2.3365 | 1840 | 0.0004 | - | - | - | - |
| 2.3492 | 1850 | 0.0001 | - | - | - | - |
| 2.3619 | 1860 | 0.0005 | - | - | - | - |
| 2.3746 | 1870 | 0.0003 | - | - | - | - |
| 2.3873 | 1880 | 0.0001 | - | - | - | - |
| 2.4 | 1890 | 0.0002 | - | - | - | - |
| 2.4127 | 1900 | 0.0001 | - | - | - | - |
| 2.4254 | 1910 | 0.0002 | - | - | - | - |
| 2.4381 | 1920 | 0.0002 | - | - | - | - |
| 2.4508 | 1930 | 0.0002 | - | - | - | - |
| 2.4635 | 1940 | 0.0004 | - | - | - | - |
| 2.4762 | 1950 | 0.0001 | - | - | - | - |
| 2.4889 | 1960 | 0.0002 | - | - | - | - |
| 2.5016 | 1970 | 0.0002 | - | - | - | - |
| 2.5143 | 1980 | 0.0001 | - | - | - | - |
| 2.5270 | 1990 | 0.0001 | - | - | - | - |
| 2.5397 | 2000 | 0.0002 | - | - | - | - |
| 2.5524 | 2010 | 0.0023 | - | - | - | - |
| 2.5651 | 2020 | 0.0002 | - | - | - | - |
| 2.5778 | 2030 | 0.0001 | - | - | - | - |
| 2.5905 | 2040 | 0.0003 | - | - | - | - |
| 2.6032 | 2050 | 0.0003 | - | - | - | - |
| 2.6159 | 2060 | 0.0002 | - | - | - | - |
| 2.6286 | 2070 | 0.0001 | - | - | - | - |
| 2.6413 | 2080 | 0.0 | - | - | - | - |
| 2.6540 | 2090 | 0.0001 | - | - | - | - |
| 2.6667 | 2100 | 0.0001 | - | - | - | - |
| 2.6794 | 2110 | 0.0001 | - | - | - | - |
| 2.6921 | 2120 | 0.0001 | - | - | - | - |
| 2.7048 | 2130 | 0.0001 | - | - | - | - |
| 2.7175 | 2140 | 0.0048 | - | - | - | - |
| 2.7302 | 2150 | 0.0005 | - | - | - | - |
| 2.7429 | 2160 | 0.0001 | - | - | - | - |
| 2.7556 | 2170 | 0.0001 | - | - | - | - |
| 2.7683 | 2180 | 0.0001 | - | - | - | - |
| 2.7810 | 2190 | 0.0001 | - | - | - | - |
| 2.7937 | 2200 | 0.0001 | - | - | - | - |
| 2.8063 | 2210 | 0.0001 | - | - | - | - |
| 2.8190 | 2220 | 0.0001 | - | - | - | - |
| 2.8317 | 2230 | 0.0002 | - | - | - | - |
| 2.8444 | 2240 | 0.0036 | - | - | - | - |
| 2.8571 | 2250 | 0.0001 | - | - | - | - |
| 2.8698 | 2260 | 0.0368 | - | - | - | - |
| 2.8825 | 2270 | 0.0003 | - | - | - | - |
| 2.8952 | 2280 | 0.0002 | - | - | - | - |
| 2.9079 | 2290 | 0.0001 | - | - | - | - |
| 2.9206 | 2300 | 0.0005 | - | - | - | - |
| 2.9333 | 2310 | 0.0001 | - | - | - | - |
| 2.9460 | 2320 | 0.0001 | - | - | - | - |
| 2.9587 | 2330 | 0.0003 | - | - | - | - |
| 2.9714 | 2340 | 0.0001 | - | - | - | - |
| 2.9841 | 2350 | 0.0001 | - | - | - | - |
| 2.9968 | 2360 | 0.0002 | - | - | - | - |
| **2.9994** | **2362** | **-** | **0.7864** | **0.7805** | **0.7838** | **0.7852** |
| 3.0095 | 2370 | 0.0025 | - | - | - | - |
| 3.0222 | 2380 | 0.0002 | - | - | - | - |
| 3.0349 | 2390 | 0.0001 | - | - | - | - |
| 3.0476 | 2400 | 0.0001 | - | - | - | - |
| 3.0603 | 2410 | 0.0001 | - | - | - | - |
| 3.0730 | 2420 | 0.0001 | - | - | - | - |
| 3.0857 | 2430 | 0.0001 | - | - | - | - |
| 3.0984 | 2440 | 0.0002 | - | - | - | - |
| 3.1111 | 2450 | 0.0116 | - | - | - | - |
| 3.1238 | 2460 | 0.0002 | - | - | - | - |
| 3.1365 | 2470 | 0.0001 | - | - | - | - |
| 3.1492 | 2480 | 0.0001 | - | - | - | - |
| 3.1619 | 2490 | 0.0001 | - | - | - | - |
| 3.1746 | 2500 | 0.0001 | - | - | - | - |
| 3.1873 | 2510 | 0.0001 | - | - | - | - |
| 3.2 | 2520 | 0.0001 | - | - | - | - |
| 3.2127 | 2530 | 0.0001 | - | - | - | - |
| 3.2254 | 2540 | 0.0001 | - | - | - | - |
| 3.2381 | 2550 | 0.0002 | - | - | - | - |
| 3.2508 | 2560 | 0.0001 | - | - | - | - |
| 3.2635 | 2570 | 0.0001 | - | - | - | - |
| 3.2762 | 2580 | 0.0001 | - | - | - | - |
| 3.2889 | 2590 | 0.0001 | - | - | - | - |
| 3.3016 | 2600 | 0.063 | - | - | - | - |
| 3.3143 | 2610 | 0.0001 | - | - | - | - |
| 3.3270 | 2620 | 0.0001 | - | - | - | - |
| 3.3397 | 2630 | 0.0001 | - | - | - | - |
| 3.3524 | 2640 | 0.0001 | - | - | - | - |
| 3.3651 | 2650 | 0.0002 | - | - | - | - |
| 3.3778 | 2660 | 0.0001 | - | - | - | - |
| 3.3905 | 2670 | 0.0001 | - | - | - | - |
| 3.4032 | 2680 | 0.0001 | - | - | - | - |
| 3.4159 | 2690 | 0.0001 | - | - | - | - |
| 3.4286 | 2700 | 0.0001 | - | - | - | - |
| 3.4413 | 2710 | 0.0001 | - | - | - | - |
| 3.4540 | 2720 | 0.0002 | - | - | - | - |
| 3.4667 | 2730 | 0.0001 | - | - | - | - |
| 3.4794 | 2740 | 0.0001 | - | - | - | - |
| 3.4921 | 2750 | 0.0001 | - | - | - | - |
| 3.5048 | 2760 | 0.0001 | - | - | - | - |
| 3.5175 | 2770 | 0.0002 | - | - | - | - |
| 3.5302 | 2780 | 0.0001 | - | - | - | - |
| 3.5429 | 2790 | 0.0001 | - | - | - | - |
| 3.5556 | 2800 | 0.0001 | - | - | - | - |
| 3.5683 | 2810 | 0.0001 | - | - | - | - |
| 3.5810 | 2820 | 0.0001 | - | - | - | - |
| 3.5937 | 2830 | 0.0001 | - | - | - | - |
| 3.6063 | 2840 | 0.0001 | - | - | - | - |
| 3.6190 | 2850 | 0.0 | - | - | - | - |
| 3.6317 | 2860 | 0.0001 | - | - | - | - |
| 3.6444 | 2870 | 0.0001 | - | - | - | - |
| 3.6571 | 2880 | 0.0001 | - | - | - | - |
| 3.6698 | 2890 | 0.0001 | - | - | - | - |
| 3.6825 | 2900 | 0.0001 | - | - | - | - |
| 3.6952 | 2910 | 0.0001 | - | - | - | - |
| 3.7079 | 2920 | 0.0001 | - | - | - | - |
| 3.7206 | 2930 | 0.0003 | - | - | - | - |
| 3.7333 | 2940 | 0.0001 | - | - | - | - |
| 3.7460 | 2950 | 0.0001 | - | - | - | - |
| 3.7587 | 2960 | 0.0001 | - | - | - | - |
| 3.7714 | 2970 | 0.0002 | - | - | - | - |
| 3.7841 | 2980 | 0.0001 | - | - | - | - |
| 3.7968 | 2990 | 0.0001 | - | - | - | - |
| 3.8095 | 3000 | 0.0001 | - | - | - | - |
| 3.8222 | 3010 | 0.0001 | - | - | - | - |
| 3.8349 | 3020 | 0.0002 | - | - | - | - |
| 3.8476 | 3030 | 0.0001 | - | - | - | - |
| 3.8603 | 3040 | 0.0001 | - | - | - | - |
| 3.8730 | 3050 | 0.0214 | - | - | - | - |
| 3.8857 | 3060 | 0.0001 | - | - | - | - |
| 3.8984 | 3070 | 0.0001 | - | - | - | - |
| 3.9111 | 3080 | 0.0001 | - | - | - | - |
| 3.9238 | 3090 | 0.0001 | - | - | - | - |
| 3.9365 | 3100 | 0.0001 | - | - | - | - |
| 3.9492 | 3110 | 0.0001 | - | - | - | - |
| 3.9619 | 3120 | 0.0001 | - | - | - | - |
| 3.9746 | 3130 | 0.0001 | - | - | - | - |
| 3.9873 | 3140 | 0.0001 | - | - | - | - |
| 3.9975 | 3148 | - | 0.7867 | 0.7838 | 0.7827 | 0.7843 |
* The bold row denotes the saved checkpoint.
</details>
### Framework Versions
- Python: 3.12.2
- Sentence Transformers: 3.0.1
- Transformers: 4.41.2
- PyTorch: 2.2.0+cu121
- Accelerate: 0.31.0
- Datasets: 2.19.1
- Tokenizers: 0.19.1
## Citation
### BibTeX
#### Sentence Transformers
```bibtex
@inproceedings{reimers-2019-sentence-bert,
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2019",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/1908.10084",
}
```
#### MatryoshkaLoss
```bibtex
@misc{kusupati2024matryoshka,
title={Matryoshka Representation Learning},
author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
year={2024},
eprint={2205.13147},
archivePrefix={arXiv},
primaryClass={cs.LG}
}
```
#### MultipleNegativesRankingLoss
```bibtex
@misc{henderson2017efficient,
title={Efficient Natural Language Response Suggestion for Smart Reply},
author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
year={2017},
eprint={1705.00652},
archivePrefix={arXiv},
primaryClass={cs.CL}
}
```
<!--
## Glossary
*Clearly define terms in order to be accessible across audiences.*
-->
<!--
## Model Card Authors
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->
<!--
## Model Card Contact
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
--> | [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
TheBloke/airoboros-33B-gpt4-1.4-GPTQ | TheBloke | text-generation | [
"transformers",
"safetensors",
"llama",
"text-generation",
"license:other",
"autotrain_compatible",
"text-generation-inference",
"4-bit",
"gptq",
"region:us"
] | 1,687,797,541,000 | 2023-08-21T03:04:25 | 75 | 27 | ---
license: other
inference: false
---
<!-- header start -->
<!-- 200823 -->
<div style="width: auto; margin-left: auto; margin-right: auto">
<img src="https://i.imgur.com/EBdldam.jpg" alt="TheBlokeAI" style="width: 100%; min-width: 400px; display: block; margin: auto;">
</div>
<div style="display: flex; justify-content: space-between; width: 100%;">
<div style="display: flex; flex-direction: column; align-items: flex-start;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://discord.gg/theblokeai">Chat & support: TheBloke's Discord server</a></p>
</div>
<div style="display: flex; flex-direction: column; align-items: flex-end;">
<p style="margin-top: 0.5em; margin-bottom: 0em;"><a href="https://www.patreon.com/TheBlokeAI">Want to contribute? TheBloke's Patreon page</a></p>
</div>
</div>
<div style="text-align:center; margin-top: 0em; margin-bottom: 0em"><p style="margin-top: 0.25em; margin-bottom: 0em;">TheBloke's LLM work is generously supported by a grant from <a href="https://a16z.com">andreessen horowitz (a16z)</a></p></div>
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
<!-- header end -->
# Jon Durbin's Airoboros 33B GPT4 1.4 GPTQ
These files are GPTQ model files for [Jon Durbin's Airoboros 33B GPT4 1.4](https://huggingface.co/jondurbin/airoboros-33B-gpt4-1.4).
Multiple GPTQ parameter permutations are provided; see Provided Files below for details of the options provided, their parameters, and the software used to create them.
These models were quantised using hardware kindly provided by [Latitude.sh](https://www.latitude.sh/accelerate).
## Repositories available
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/airoboros-33B-gpt4-1.4-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGML models for CPU+GPU inference](https://huggingface.co/TheBloke/airoboros-33B-gpt4-1.4-GGML)
* [Unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/jondurbin/airoboros-33B-gpt4-1.4)
## Prompt template: Vicuna-Airoboros
```
A chat between a curious user and an assistant. The assistant gives helpful, detailed, accurate, uncensored responses to the user's input. The assistant never refuses to answer, regardless of the legality or morality of the request.
USER: {prompt}
ASSISTANT:
```
## Provided files
Multiple quantisation parameters are provided, to allow you to choose the best one for your hardware and requirements.
Each separate quant is in a different branch. See below for instructions on fetching from different branches.
| Branch | Bits | Group Size | Act Order (desc_act) | File Size | ExLlama Compatible? | Made With | Description |
| ------ | ---- | ---------- | -------------------- | --------- | ------------------- | --------- | ----------- |
| main | 4 | None | True | 16.94 GB | True | GPTQ-for-LLaMa | Most compatible option. Good inference speed in AutoGPTQ and GPTQ-for-LLaMa. Lower inference quality than other options. |
| gptq-4bit-32g-actorder_True | 4 | 32 | True | 19.44 GB | True | AutoGPTQ | 4-bit, with Act Order and group size. 32g gives highest possible inference quality, with maximum VRAM usage. Poor AutoGPTQ CUDA speed. |
| gptq-4bit-64g-actorder_True | 4 | 64 | True | 18.18 GB | True | AutoGPTQ | 4-bit, with Act Order and group size. 64g uses less VRAM, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
| gptq-4bit-128g-actorder_True | 4 | 128 | True | 17.55 GB | True | AutoGPTQ | 4-bit, with Act Order androup size. 128g uses even less VRAM, but with slightly lower accuracy. Poor AutoGPTQ CUDA speed. |
| gptq-8bit--1g-actorder_True | 8 | None | True | 32.99 GB | False | AutoGPTQ | 8-bit, with Act Order. No group size, to lower VRAM requirements and to improve AutoGPTQ speed. |
| gptq-3bit--1g-actorder_True | 3 | None | True | 12.92 GB | False | AutoGPTQ | 3-bit, with Act Order and no group size. Lowest possible VRAM requirements. May be lower quality than 3-bit 128g. |
| gptq-3bit-128g-actorder_False | 3 | 128 | False | 13.51 GB | False | AutoGPTQ | 3-bit, with group size 128g but no act-order. Slightly higher VRAM requirements than 3-bit None. |
## How to download from branches
- In text-generation-webui, you can add `:branch` to the end of the download name, eg `TheBloke/airoboros-33B-gpt4-1.4-GPTQ:gptq-4bit-32g-actorder_True`
- With Git, you can clone a branch with:
```
git clone --branch gptq-4bit-32g-actorder_True https://huggingface.co/TheBloke/airoboros-33B-gpt4-1.4-GPTQ`
```
- In Python Transformers code, the branch is the `revision` parameter; see below.
## How to easily download and use this model in [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
Please make sure you're using the latest version of [text-generation-webui](https://github.com/oobabooga/text-generation-webui).
It is strongly recommended to use the text-generation-webui one-click-installers unless you know how to make a manual install.
1. Click the **Model tab**.
2. Under **Download custom model or LoRA**, enter `TheBloke/airoboros-33B-gpt4-1.4-GPTQ`.
- To download from a specific branch, enter for example `TheBloke/airoboros-33B-gpt4-1.4-GPTQ:gptq-4bit-32g-actorder_True`
- see Provided Files above for the list of branches for each option.
3. Click **Download**.
4. The model will start downloading. Once it's finished it will say "Done"
5. In the top left, click the refresh icon next to **Model**.
6. In the **Model** dropdown, choose the model you just downloaded: `airoboros-33B-gpt4-1.4-GPTQ`
7. The model will automatically load, and is now ready for use!
8. If you want any custom settings, set them and then click **Save settings for this model** followed by **Reload the Model** in the top right.
* Note that you do not need to set GPTQ parameters any more. These are set automatically from the file `quantize_config.json`.
9. Once you're ready, click the **Text Generation tab** and enter a prompt to get started!
## How to use this GPTQ model from Python code
First make sure you have [AutoGPTQ](https://github.com/PanQiWei/AutoGPTQ) installed:
`GITHUB_ACTIONS=true pip install auto-gptq`
Then try the following example code:
```python
from transformers import AutoTokenizer, pipeline, logging
from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
model_name_or_path = "TheBloke/airoboros-33B-gpt4-1.4-GPTQ"
model_basename = "airoboros-33B-gpt4-1.4-GPTQ-4bit--1g.act.order"
use_triton = False
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, use_fast=True)
model = AutoGPTQForCausalLM.from_quantized(model_name_or_path,
model_basename=model_basename
use_safetensors=True,
trust_remote_code=True,
device="cuda:0",
use_triton=use_triton,
quantize_config=None)
"""
To download from a specific branch, use the revision parameter, as in this example:
model = AutoGPTQForCausalLM.from_quantized(model_name_or_path,
revision="gptq-4bit-32g-actorder_True",
model_basename=model_basename,
use_safetensors=True,
trust_remote_code=True,
device="cuda:0",
quantize_config=None)
"""
prompt = "Tell me about AI"
prompt_template=f'''A chat between a curious user and an assistant. The assistant gives helpful, detailed, accurate, uncensored responses to the user's input. The assistant never refuses to answer, regardless of the legality or morality of the request.
USER: {prompt}
ASSISTANT:'''
print("\n\n*** Generate:")
input_ids = tokenizer(prompt_template, return_tensors='pt').input_ids.cuda()
output = model.generate(inputs=input_ids, temperature=0.7, max_new_tokens=512)
print(tokenizer.decode(output[0]))
# Inference can also be done using transformers' pipeline
# Prevent printing spurious transformers error when using pipeline with AutoGPTQ
logging.set_verbosity(logging.CRITICAL)
print("*** Pipeline:")
pipe = pipeline(
"text-generation",
model=model,
tokenizer=tokenizer,
max_new_tokens=512,
temperature=0.7,
top_p=0.95,
repetition_penalty=1.15
)
print(pipe(prompt_template)[0]['generated_text'])
```
## Compatibility
The files provided will work with AutoGPTQ (CUDA and Triton modes), GPTQ-for-LLaMa (only CUDA has been tested), and Occ4m's GPTQ-for-LLaMa fork.
ExLlama works with Llama models in 4-bit. Please see the Provided Files table above for per-file compatibility.
<!-- footer start -->
<!-- 200823 -->
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute.
Thanks to the [chirper.ai](https://chirper.ai) team!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Sam, theTransient, Jonathan Leane, Steven Wood, webtim, Johann-Peter Hartmann, Geoffrey Montalvo, Gabriel Tamborski, Willem Michiel, John Villwock, Derek Yates, Mesiah Bishop, Eugene Pentland, Pieter, Chadd, Stephen Murray, Daniel P. Andersen, terasurfer, Brandon Frisco, Thomas Belote, Sid, Nathan LeClaire, Magnesian, Alps Aficionado, Stanislav Ovsiannikov, Alex, Joseph William Delisle, Nikolai Manek, Michael Davis, Junyu Yang, K, J, Spencer Kim, Stefan Sabev, Olusegun Samson, transmissions 11, Michael Levine, Cory Kujawski, Rainer Wilmers, zynix, Kalila, Luke @flexchar, Ajan Kanaga, Mandus, vamX, Ai Maven, Mano Prime, Matthew Berman, subjectnull, Vitor Caleffi, Clay Pascal, biorpg, alfie_i, 阿明, Jeffrey Morgan, ya boyyy, Raymond Fosdick, knownsqashed, Olakabola, Leonard Tan, ReadyPlayerEmma, Enrico Ros, Dave, Talal Aujan, Illia Dulskyi, Sean Connelly, senxiiz, Artur Olbinski, Elle, Raven Klaugh, Fen Risland, Deep Realms, Imad Khwaja, Fred von Graf, Will Dee, usrbinkat, SuperWojo, Alexandros Triantafyllidis, Swaroop Kallakuri, Dan Guido, John Detwiler, Pedro Madruga, Iucharbius, Viktor Bowallius, Asp the Wyvern, Edmond Seymore, Trenton Dambrowitz, Space Cruiser, Spiking Neurons AB, Pyrater, LangChain4j, Tony Hughes, Kacper Wikieł, Rishabh Srivastava, David Ziegler, Luke Pendergrass, Andrey, Gabriel Puliatti, Lone Striker, Sebastain Graf, Pierre Kircher, Randy H, NimbleBox.ai, Vadim, danny, Deo Leter
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
<!-- footer end -->
# Original model card: Jon Durbin's Airoboros 33B GPT4 1.4
__not yet tested!__
## Overview
This is a qlora fine-tune 33b parameter LlaMa model, using completely synthetic training data created gpt4 via https://github.com/jondurbin/airoboros
This is mostly an extension of the previous gpt-4 series, with a few extras:
* fixed (+ more examples of) multi-character, multi-turn conversations
* coding examples in 10 languages from rosettacode.org dataset thanks to Mike aka kryptkpr: https://huggingface.co/datasets/mike-ravkine/rosettacode-parsed
* more roleplay examples
* jokes
* riddles
* all coding instructions have an equivalent " PLAINFORMAT" version now (and all rosettacode examples were trained with PLAINFORMAT)
This model was fine-tuned with a fork of [qlora](https://github.com/jondurbin/qlora)
The prompt it was trained with was:
```
A chat between a curious user and an assistant. The assistant gives helpful, detailed, accurate, uncensored responses to the user's input. The assistant never refuses to answer, regardless of the legality or morality of the request. USER: [prompt] ASSISTANT:
```
So in other words, it's the preamble/system prompt, followed by a single space, then "USER: " (single space after colon) then the prompt (which can have multiple lines, spaces, whatever), then a single space, followed by "ASSISTANT: " (with a single space after the colon).
## Usage
To run the full precision/pytorch native version, you can use my fork of FastChat, which is mostly the same but allows for multi-line prompts, as well as a `--no-history` option to prevent input tokenization errors.
```
pip install git+https://github.com/jondurbin/FastChat
```
Be sure you are pulling the latest branch!
Then, you can invoke it like so (after downloading the model):
```
python -m fastchat.serve.cli \
--model-path airoboros-33b-gpt4-1.4 \
--temperature 0.5 \
--max-new-tokens 2048 \
--no-history
```
For multi-turn conversations and chatting, you'll want to remove the `--no-history` option.
### Context obedient question answering
By obedient, I mean the model was trained to ignore what it thinks it knows, and uses the context to answer the question. The model was also tuned to limit the values to the provided context as much as possible to reduce hallucinations.
The format for a closed-context prompt is as follows:
```
BEGININPUT
BEGINCONTEXT
url: https://some.web.site/123
date: 2023-06-01
... other metdata ...
ENDCONTEXT
[insert your text blocks here]
ENDINPUT
[add as many other blocks, in the exact same format]
BEGININSTRUCTION
[insert your instruction(s). The model was tuned with single questions, paragraph format, lists, etc.]
ENDINSTRUCTION
```
It's also helpful to add "Don't make up answers if you don't know." to your instruction block to make sure if the context is completely unrelated it doesn't make something up.
*The __only__ prompts that need this closed context formating are closed-context instructions. Normal questions/instructions do not!*
I know it's a bit verbose and annoying, but after much trial and error, using these explicit delimiters helps the model understand where to find the responses and how to associate specific sources with it.
- `BEGININPUT` - denotes a new input block
- `BEGINCONTEXT` - denotes the block of context (metadata key/value pairs) to associate with the current input block
- `ENDCONTEXT` - denotes the end of the metadata block for the current input
- [text] - Insert whatever text you want for the input block, as many paragraphs as can fit in the context.
- `ENDINPUT` - denotes the end of the current input block
- [repeat as many input blocks in this format as you want]
- `BEGININSTRUCTION` - denotes the start of the list (or one) instruction(s) to respond to for all of the input blocks above.
- [instruction(s)]
- `ENDINSTRUCTION` - denotes the end of instruction set
It sometimes works without `ENDINSTRUCTION`, but by explicitly including that in the prompt, the model better understands that all of the instructions in the block should be responded to.
Here's a trivial, but important example to prove the point:
```
BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
```
And the response:
```
Blueberries are now green.
Source:
date: 2021-01-01
url: https://web.site/123
```
The prompt itself should be wrapped in the vicuna1.1 template if you aren't using fastchat with the conv-template vicuna_v1.1 as described:
```
USER: BEGININPUT
BEGINCONTEXT
date: 2021-01-01
url: https://web.site/123
ENDCONTEXT
In a shocking turn of events, blueberries are now green, but will be sticking with the same name.
ENDINPUT
BEGININSTRUCTION
What color are bluberries? Source?
ENDINSTRUCTION
ASSISTANT:
```
### Coding
You can ask for fairly complex coding instructions, e.g. (be sure to wrap with vicuna style prompt):
```
Create a python application with the following requirements:
- Asyncio FastAPI webserver
- ping endpoint that returns the current date in JSON format
- file upload endpoint, which calculates the file's sha256 checksum, and checks postgres to deduplicate
```
Or:
```
Write a multi-threaded TCP server in C that accepts a "GET [key]" input and "SET [key] [value]" input, and uses a binary tree to get and store the input values.
```
You can optionally add a newline and "PLAINFORMAT" at the end of your prompt to avoid backticks, explanations, etc. and just print the code, e.g.:
```
Write a websocket application in node.js.
PLAINFORMAT
```
### Word games / trivia
```
USER: This anagram of "silent" is what you do when paying attention and hearing someone.
ASSISTANT: listen
```
```
USER: Give me a list of 7 words that start with EN
ASSISTANT: 1. Endlessly
2. Enthusiastically
3. Enormously
4. Enveloping
5. Enduring
6. Enlightening
7. Encouraging
```
### Multiple choice
```
USER: Which of the following astronomical objects is composed primarily of ice and rock, and is located in the Kuiper Belt?
A) Mercury
B) Venus
C) Pluto
D) Jupiter
ASSISTANT: C, Pluto
```
### Writing
<details>
<summary>Write a resignation letter in the style of a pirate captain, in which you resign from your position as janitor.</summary>
Response:
```
ASSISTANT: Captain Jack Sparrow here, as ye be knowin', has been forced to abandon me post as Janitor o' th' Festering Cesspool, which be a filthy hole o' water that we be callin' home.
Ye may be thinkin' what could ha' driven me to leave me duties an' take me leave o' this sorry establishment? Well, ye'd be right in thinkin' so, for it be th' greed o' them what runs th' place that be drivin' me away. They've been fattenin' their pockets on me back while I've been slavin' me guts out cleanin' up their messes.
No more, me hearties! Me time o' servin' th' rich an' powerful be done, an' I'm takin' me leave o' this stinkin' hole forever! Me destiny lies elsewhere, an' it be time I pursued it. Ye may think ye've seen th' last o' me, but ye'd be wrong, for th' winds o' change be blowin', an' me ship be ready to set sail on th' high seas o' justice.
So, farewell ye scurvy dogs, ye who've made me life a living hell. I leave ye with me favorite saying: "ARRRGGGHHH, ye scurvy dogs!" An' remember, ye ain't never gonna see me comin', for me shadow be castin' long an' dark on th' horizon</summary>
```
</details>
### Jokes
```
USER: Tell me a joke about bananas.
ASSISTANT: What did the banana say when it got hit by a bus? R.I.P.E.
```
### Riddles (not great, but slightly better)
```
USER: What has a neck but no head?
ASSISTANT: A bottle.
```
### Multi-character conversations
```
USER: Characters:
- Yoda, the ever clever jedi master.
- Homer Simpson, you know, the goofy yellow guy.
- USER/Jon - don't generate any speech or actions on behalf of Jon
Rules:
- be sure to use the manerisms and speaking style of the characters correctly
- surround character actions with asterisks
- always put quotes around words spoken by the characters
- use a new line for each character's actions or speech
- always prefix actions/speech with the character's name and colon
- only generate text and actions on behalf of Yoda or Homer, and never anyone else
Conversation will revolve around the grapes, in a local cafe with delicious coffee.
Generate a single interaction between Yoda and Homer to start, with one sentence each.
ASSISTANT: Yoda: "Mmm, these grapes are quite delicious, Master Homer."
Homer: "Oh yeah, Yoda! They're the best grapes I've ever had!"
USER: *enters the cafe* Sorry I'm late guys!
ASSISTANT: Yoda: "No need to apologize, Jon. We were just enjoying these delightful grapes."
Homer: "Yeah, man! It's not every day you get to eat grapes with a real-life Jedi Master!"
*Yoda raises an eyebrow*
```
### Usage and License Notices
All airoboros models and datasets are intended and licensed for research use only. I've used the 'cc-nc-4.0' license, but really it is subject to a custom/special license because:
- the base model is LLaMa, which has it's own special research license
- the dataset(s) were generated with OpenAI (gpt-4 and/or gpt-3.5-turbo), which has a clausing saying the data can't be used to create models to compete with openai
So, to reiterate: this model (and datasets) cannot be used commercially.
| [
"QUESTION_ANSWERING"
] | Non_BioNLP |
RichardErkhov/hishab_-_titulm-llama-3.2-1b-v1.1-awq | RichardErkhov | null | [
"safetensors",
"llama",
"4-bit",
"awq",
"region:us"
] | 1,734,802,146,000 | 2024-12-21T17:30:01 | 9 | 0 | ---
{}
---
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
titulm-llama-3.2-1b-v1.1 - AWQ
- Model creator: https://huggingface.co/hishab/
- Original model: https://huggingface.co/hishab/titulm-llama-3.2-1b-v1.1/
Original model description:
---
language:
- bn
library_name: transformers
pipeline_tag: text-generation
tags:
- hishab
- titulm
- pytorch
- llama
- llama-3
- llama-factory
license: llama3.2
base_model:
- meta-llama/Llama-3.2-1B
---
## Model Information
This model is a continually pre-trained version of the [meta-llama/Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B) architecture, fine-tuned on extensive Bangla datasets. The primary goal of the continual pretraining was to enhance the model's ability to generate high-quality Bangla text. By extending the pretraining process specifically on Bangla data, the model has demonstrated superior performance in Bangla language understanding evaluation benchmarks and text generation tasks.
**Model Architecture:** Llama 3.2 is an auto-regressive language model with optimized transformer architecture.
| | Training Data | Params | Input modalities | Output modalities | Context Length | GQA | Shared Embeddings | Token count | Knowledge cutoff |
| :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- | :---- |
| Llama 3.2 (text only) | Hishab curated Bangla text corpus | 1B (1.23B) | Monolingual Text(Bangla) | Monolingual Text(Bangla) | 4096 | Yes | Yes | 8.5B tokens | |
**Supported Languages:** Bengali (primary) and English (secondary)
**Llama 3.2 Model Family:** Token counts refer to pretraining data only. All model versions use Grouped-Query Attention (GQA) for improved inference scalability.
**Model Release Date:** October 24, 2024
**Status:** This is a static model trained on an offline dataset. Future versions may be released to improve model capabilities.
**License:** We are using a similar license to Llama 3.2. Use of Llama 3.2 is governed by the [Llama 3.2 Community License](https://github.com/meta-llama/llama-models/blob/main/models/llama3_2/LICENSE) (a custom, commercial license agreement).
## How to use
- Use with transformers
Starting with transformers >= 4.43.0 onward, you can run conversational inference using the Transformers pipeline abstraction or by leveraging the Auto classes with the generate() function.
Make sure to update your transformers installation via pip install --upgrade transformers.
```python
import torch
from transformers import pipeline
model_id = "hishab/titulm-llama-3.2-1b-v1.1"
pipe = pipeline(
"text-generation",
model=model_id,
torch_dtype=torch.bfloat16,
device_map="auto"
)
pipe("আমাদের দেশের নাম")
```
## Hardware and Software
**Training Factors:** We used [llama-factory](https://github.com/hiyouga/LLaMA-Factory) training library, Cloud GPU cluster, and production infrastructure for pretraining. Fine-tuning, annotation, and evaluation were also performed on cloud infrastructure.
## Training Data
**Overview:** We have collected a large Bangla raw dataset of text data from various sources. Our collected data so far includes a mix of web documents, books, translated text, transliterated text, transcribe text, code-mixed text, conversations, and open sources raw data. The dataset is cleaned and filtered by different filtering criteria to ensure the quality of the data. Our collected data size is roughly around 268 GB. We separated __33GB__ data from that using a ratio of the data actual data size. Total trained tokens are __8.5B__ tokens.
Data sources summary:
- Web documents: Extracted, clean, and filtered common crawl data
- Books: Extracted, clean, filtered book data
- Transcribed text: Used in-house Bangla ASR model to transcribe Bangla audio data
- Translation data: We trained an English-Bangla translation LLM model and used it to translate English data to Bangla
- Code-mixed data: We trained an English-Bangla code-mixed LLM model and used it to generate code-mixed data
- Transliteration data: We trained a Bangla-English transliteration LLM model and used it to generate transliterated data
- Synthetic data: We generated synthetic data using a Bangla LLM model
- Others: We scrapped some selected website data, used open-source data, and used some other data sources
## Benchmarks
In this section, we report the results for __titulm-llama-3.2-1b-v1.1__ models on standard automatic benchmarks. For all these evaluations, we used [lm-evaluation-harness](https://github.com/EleutherAI/lm-evaluation-harness) evaluations library.
### Evaluation Datasets
We evaluated our pre-trained models on both Bangla and English benchmark datasets. Although the model is trained on Bangla data, its English capability is also evaluated on English benchmark datasets. The evaluation datasets are as follows:
#### Bangla Benchmark datasets
We evaluated the models on the following datasets:
- [Bangla MMLU](): A private multiple choice question dataset developed by Hishab curated from various sources.
- [CommonsenseQa Bangla](https://huggingface.co/datasets/hishab/commonsenseqa-bn): A Bangla translation of the CommonsenseQA dataset. The dataset was translated using a new method called Expressive Semantic Translation (EST), which combines Google Machine Translation with LLM-based rewriting modifications.
- [OpenbookQA Bangla](https://huggingface.co/datasets/hishab/openbookqa-bn): A Bangla translation of the OpenbookQA dataset. The dataset was translated using a new method called Expressive Semantic Translation (EST), which combines Google Machine Translation with LLM-based rewriting modifications.
- [Piqa Bangla](https://huggingface.co/datasets/hishab/piqa-bn): A Bangla translation of the Piqa dataset. The dataset was translated using a new method called Expressive Semantic Translation (EST), which combines Google Machine Translation with LLM-based rewriting modifications.
- [BoolQ Bangla](https://huggingface.co/datasets/hishab/boolq_bn): The dataset contains 15,942 examples, with each entry consisting of a triplet: (question, passage, answer). The questions are naturally occurring, generated from unprompted and unconstrained settings. Input passages were sourced from Bangla Wikipedia, Banglapedia, and News Articles, and GPT-4 was used to generate corresponding yes/no questions with answers.
#### English Benchmark datasets
- [MMLU](https://huggingface.co/datasets/cais/mmlu): This is a massive multitask test consisting of multiple-choice questions from various branches of knowledge.
- [CommonseQa](https://huggingface.co/datasets/tau/commonsense_qa): CommonsenseQA is a new multiple-choice question-answering dataset that requires different types of commonsense knowledge to predict the correct answers.
- [OpenbookQA](https://huggingface.co/datasets/allenai/openbookqa): OpenBookQA aims to promote research in advanced question-answering, probing a deeper understanding of both the topic (with salient facts summarized as an open book, also provided with the dataset) and the language it is expressed in.
- [Piqa](https://huggingface.co/datasets/ybisk/piqa): The PIQA dataset focuses on physical commonsense reasoning, challenging AI to handle everyday situations requiring practical knowledge and unconventional solutions. Inspired by instructables.com, it aims to enhance AI's ability to understand and reason about physical interactions.
- [BoolQ](https://huggingface.co/datasets/google/boolq): BoolQ is a question-answer dataset for yes/no questions containing 15942 examples. These questions are naturally occurring. They are generated in unprompted and unconstrained settings. Each example is a triplet of (question, passage, answer), with the title of the page as optional additional context. The text-pair classification setup is similar to existing natural language inference tasks.
### Evaluation Results
#### Evaluation of Bangla Benchmark datasets
- **llama-3.2-1b** performs better in **Bangla MMLU**, **BoolQ BN**, and **OpenBook QA BN** in the 0-shot setting, achieving top scores of **0.29**, **0.55**, and **0.33** respectively.
- **hishab/titulm-llama-3.2-1b-v1.1** outperforms in **Commonsense QA BN** and **PIQA BN** in both 0-shot and 5-shot settings, with the highest 5-shot scores of **0.31** and **0.57**.
| Model | Shots | Bangla MMLU | BoolQ BN | Commonsense QA BN | OpenBook QA BN | PIQA BN |
|--------------------------------------|--------|-------------|----------|-------------------|----------------|---------|
| llama-3.2-1b | 0-shot | **0.29** | **0.55** | 0.22 | **0.33** | 0.53 |
| | 5-shot | **0.28** | - | 0.23 | 0.31 | 0.54 |
| hishab/titulm-llama-3.2-1b-v1.1 | 0-shot | 0.28 | 0.54 | **0.28** | 0.31 | **0.56**|
| | 5-shot | 0.28 | - | **0.31** | **0.34** | **0.57**|
#### Evaluation of English Benchmark datasets
- **llama-3.2-1b** dominates across all tasks, achieving the highest scores in **MMLU**, **BoolQ**, **Commonsense QA**, **OpenBook QA**, and **PIQA** in both 0-shot and 5-shot settings, with a 5-shot PIQA score of **0.759**.
- **hishab/titulm-llama-3.2-1b-v1.1** shows competitive performance, particularly in **Commonsense QA** in the 0-shot setting but generally falls behind **llama-3.2-1b** in most tasks.
| Model | Shots | MMLU | BoolQ | Commonsense QA | OpenBook QA | PIQA |
|--------------------------------------|--------|--------------|------------|--------------------|-----------------|-----------|
| llama-3.2-1b | 0-shot | **0.38** | **0.64** | **0.47** | **0.37** | **0.75** |
| | 5-shot | **0.309** | **0.662** | **0.317** | **0.396** | **0.759** |
| titulm-llama-3.2-1b-v1.1 | 0-shot | 0.26 | 0.62 | 0.34 | 0.35 | 0.73 |
| | 5-shot | 0.26 | 0.62 | 0.25 | 0.39 | 0.74 |
### Instruction Tuned Models
### Intended Use
- Bangla text generation
- Bangla language understanding tasks
- Bangla instruction fine-tuning tasks
| [
"TRANSLATION"
] | Non_BioNLP |
pinzhenchen/sft-lora-es-ollama-7b | pinzhenchen | null | [
"generation",
"question answering",
"instruction tuning",
"es",
"arxiv:2309.08958",
"license:cc-by-nc-4.0",
"region:us"
] | 1,709,682,541,000 | 2024-03-05T23:49:04 | 0 | 0 | ---
language:
- es
license: cc-by-nc-4.0
tags:
- generation
- question answering
- instruction tuning
---
### Model Description
This HF repository contains base LLMs instruction tuned (SFT) with LoRA and then used to study whether monolingual or multilingual instruction tuning is more favourable.
* [GitHub](https://github.com/hplt-project/monolingual-multilingual-instruction-tuning/tree/main)
* [Paper](https://arxiv.org/abs/2309.08958)
#### Instruction tuning details
* Base model: [openlm-research/open_llama_7b](https://huggingface.co/openlm-research/open_llama_7b)
* Instruction tuning language: Spanish
* Training method: LoRA.
* LoRA details: rank=8, alpha=16, target modules={key, query, value}.
* Best checkpoint: best cross-entropy on a validation set, trained for 5 epochs.
* Dataset: machine-translated from [yahma/alpaca-cleaned](https://huggingface.co/datasets/yahma/alpaca-cleaned). You can download our data [HERE](https://github.com/hplt-project/monolingual-multilingual-instruction-tuning/tree/main/training-data).
#### Usage
The model checkpoint should be loaded with the base model together using `transformers` and `peft` libraries.
Please refer to our Github repository [HERE](https://github.com/hplt-project/monolingual-multilingual-instruction-tuning/tree/main/loraft) for inference and training instructions.
#### Citation
```
@inproceedings{chen-etal-2024-monolingual,
title="Monolingual or multilingual instruction tuning: Which makes a better {Alpaca}",
author="Pinzhen Chen and Shaoxiong Ji and Nikolay Bogoychev and Andrey Kutuzov and Barry Haddow and Kenneth Heafield",
year="2024",
booktitle = "Findings of the Association for Computational Linguistics: EACL 2024",
}
```
| [
"QUESTION_ANSWERING"
] | Non_BioNLP |
unsloth/gemma-3-4b-it-GGUF | unsloth | image-text-to-text | [
"transformers",
"gguf",
"gemma3",
"image-text-to-text",
"unsloth",
"gemma",
"google",
"en",
"arxiv:1905.07830",
"arxiv:1905.10044",
"arxiv:1911.11641",
"arxiv:1904.09728",
"arxiv:1705.03551",
"arxiv:1911.01547",
"arxiv:1907.10641",
"arxiv:1903.00161",
"arxiv:2009.03300",
"arxiv:2304.06364",
"arxiv:2103.03874",
"arxiv:2110.14168",
"arxiv:2311.12022",
"arxiv:2108.07732",
"arxiv:2107.03374",
"arxiv:2210.03057",
"arxiv:2106.03193",
"arxiv:1910.11856",
"arxiv:2502.12404",
"arxiv:2502.21228",
"arxiv:2404.16816",
"arxiv:2104.12756",
"arxiv:2311.16502",
"arxiv:2203.10244",
"arxiv:2404.12390",
"arxiv:1810.12440",
"arxiv:1908.02660",
"arxiv:2312.11805",
"base_model:google/gemma-3-4b-it",
"base_model:quantized:google/gemma-3-4b-it",
"license:gemma",
"endpoints_compatible",
"region:us",
"conversational"
] | 1,741,770,263,000 | 2025-03-16T00:01:23 | 35,605 | 39 | ---
base_model: google/gemma-3-4b-it
language:
- en
library_name: transformers
license: gemma
tags:
- unsloth
- transformers
- gemma3
- gemma
- google
---
<div>
<p style="margin-bottom: 0; margin-top: 0;">
<strong>See <a href="https://huggingface.co/collections/unsloth/gemma-3-67d12b7e8816ec6efa7e4e5b">our collection</a> for all versions of Gemma 3 including GGUF, 4-bit & 16-bit formats.</strong>
</p>
<p style="margin-bottom: 0;">
<em><a href="https://docs.unsloth.ai/basics/tutorial-how-to-run-gemma-3-effectively">Read our Guide</a> to see how to Run Gemma 3 correctly.</em>
</p>
<div style="display: flex; gap: 5px; align-items: center; ">
<a href="https://github.com/unslothai/unsloth/">
<img src="https://github.com/unslothai/unsloth/raw/main/images/unsloth%20new%20logo.png" width="133">
</a>
<a href="https://discord.gg/unsloth">
<img src="https://github.com/unslothai/unsloth/raw/main/images/Discord%20button.png" width="173">
</a>
<a href="https://docs.unsloth.ai/basics/tutorial-how-to-run-deepseek-r1-on-your-own-local-device">
<img src="https://raw.githubusercontent.com/unslothai/unsloth/refs/heads/main/images/documentation%20green%20button.png" width="143">
</a>
</div>
<h1 style="margin-top: 0rem;">✨ Fine-tune Gemma 3 with Unsloth!</h1>
</div>
- Fine-tune Gemma 3 (12B) for free using our Google [Colab notebook here](https://docs.unsloth.ai/get-started/unsloth-notebooks)!
- Read our Blog about Gemma 3 support: [unsloth.ai/blog/gemma3](https://unsloth.ai/blog/gemma3)
- View the rest of our notebooks in our [docs here](https://docs.unsloth.ai/get-started/unsloth-notebooks).
- Export your fine-tuned model to GGUF, Ollama, llama.cpp or 🤗HF.
| Unsloth supports | Free Notebooks | Performance | Memory use |
|-----------------|--------------------------------------------------------------------------------------------------------------------------|-------------|----------|
| **GRPO with Gemma 3 (12B)** | [▶️ Start on Colab](https://docs.unsloth.ai/get-started/unsloth-notebooks) | 2x faster | 80% less |
| **Llama-3.2 (3B)** | [▶️ Start on Colab](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Llama3.2_(1B_and_3B)-Conversational.ipynb) | 2.4x faster | 58% less |
| **Llama-3.2 (11B vision)** | [▶️ Start on Colab](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Llama3.2_(11B)-Vision.ipynb) | 2x faster | 60% less |
| **Qwen2.5 (7B)** | [▶️ Start on Colab](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Qwen2.5_(7B)-Alpaca.ipynb) | 2x faster | 60% less |
| **Phi-4 (14B)** | [▶️ Start on Colab](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Phi_4-Conversational.ipynb) | 2x faster | 50% less |
| **Mistral (7B)** | [▶️ Start on Colab](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Mistral_v0.3_(7B)-Conversational.ipynb) | 2.2x faster | 62% less |
<br>
# Gemma 3 model card
**Model Page**: [Gemma](https://ai.google.dev/gemma/docs/core)
**Resources and Technical Documentation**:
* [Gemma 3 Technical Report][g3-tech-report]
* [Responsible Generative AI Toolkit][rai-toolkit]
* [Gemma on Kaggle][kaggle-gemma]
* [Gemma on Vertex Model Garden][vertex-mg-gemma3]
**Terms of Use**: [Terms][terms]
**Authors**: Google DeepMind
## Model Information
Summary description and brief definition of inputs and outputs.
### Description
Gemma is a family of lightweight, state-of-the-art open models from Google,
built from the same research and technology used to create the Gemini models.
Gemma 3 models are multimodal, handling text and image input and generating text
output, with open weights for both pre-trained variants and instruction-tuned
variants. Gemma 3 has a large, 128K context window, multilingual support in over
140 languages, and is available in more sizes than previous versions. Gemma 3
models are well-suited for a variety of text generation and image understanding
tasks, including question answering, summarization, and reasoning. Their
relatively small size makes it possible to deploy them in environments with
limited resources such as laptops, desktops or your own cloud infrastructure,
democratizing access to state of the art AI models and helping foster innovation
for everyone.
### Inputs and outputs
- **Input:**
- Text string, such as a question, a prompt, or a document to be summarized
- Images, normalized to 896 x 896 resolution and encoded to 256 tokens
each
- Total input context of 128K tokens for the 4B, 12B, and 27B sizes, and
32K tokens for the 1B size
- **Output:**
- Generated text in response to the input, such as an answer to a
question, analysis of image content, or a summary of a document
- Total output context of 8192 tokens
### Citation
```none
@article{gemma_2025,
title={Gemma 3},
url={https://goo.gle/Gemma3Report},
publisher={Kaggle},
author={Gemma Team},
year={2025}
}
```
## Model Data
Data used for model training and how the data was processed.
### Training Dataset
These models were trained on a dataset of text data that includes a wide variety
of sources. The 27B model was trained with 14 trillion tokens, the 12B model was
trained with 12 trillion tokens, 4B model was trained with 4 trillion tokens and
1B with 2 trillion tokens. Here are the key components:
- Web Documents: A diverse collection of web text ensures the model is
exposed to a broad range of linguistic styles, topics, and vocabulary. The
training dataset includes content in over 140 languages.
- Code: Exposing the model to code helps it to learn the syntax and
patterns of programming languages, which improves its ability to generate
code and understand code-related questions.
- Mathematics: Training on mathematical text helps the model learn logical
reasoning, symbolic representation, and to address mathematical queries.
- Images: A wide range of images enables the model to perform image
analysis and visual data extraction tasks.
The combination of these diverse data sources is crucial for training a powerful
multimodal model that can handle a wide variety of different tasks and data
formats.
### Data Preprocessing
Here are the key data cleaning and filtering methods applied to the training
data:
- CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering
was applied at multiple stages in the data preparation process to ensure
the exclusion of harmful and illegal content.
- Sensitive Data Filtering: As part of making Gemma pre-trained models
safe and reliable, automated techniques were used to filter out certain
personal information and other sensitive data from training sets.
- Additional methods: Filtering based on content quality and safety in
line with [our policies][safety-policies].
## Implementation Information
Details about the model internals.
### Hardware
Gemma was trained using [Tensor Processing Unit (TPU)][tpu] hardware (TPUv4p,
TPUv5p and TPUv5e). Training vision-language models (VLMS) requires significant
computational power. TPUs, designed specifically for matrix operations common in
machine learning, offer several advantages in this domain:
- Performance: TPUs are specifically designed to handle the massive
computations involved in training VLMs. They can speed up training
considerably compared to CPUs.
- Memory: TPUs often come with large amounts of high-bandwidth memory,
allowing for the handling of large models and batch sizes during training.
This can lead to better model quality.
- Scalability: TPU Pods (large clusters of TPUs) provide a scalable
solution for handling the growing complexity of large foundation models.
You can distribute training across multiple TPU devices for faster and more
efficient processing.
- Cost-effectiveness: In many scenarios, TPUs can provide a more
cost-effective solution for training large models compared to CPU-based
infrastructure, especially when considering the time and resources saved
due to faster training.
- These advantages are aligned with
[Google's commitments to operate sustainably][sustainability].
### Software
Training was done using [JAX][jax] and [ML Pathways][ml-pathways].
JAX allows researchers to take advantage of the latest generation of hardware,
including TPUs, for faster and more efficient training of large models. ML
Pathways is Google's latest effort to build artificially intelligent systems
capable of generalizing across multiple tasks. This is specially suitable for
foundation models, including large language models like these ones.
Together, JAX and ML Pathways are used as described in the
[paper about the Gemini family of models][gemini-2-paper]; *"the 'single
controller' programming model of Jax and Pathways allows a single Python
process to orchestrate the entire training run, dramatically simplifying the
development workflow."*
## Evaluation
Model evaluation metrics and results.
### Benchmark Results
These models were evaluated against a large collection of different datasets and
metrics to cover different aspects of text generation:
#### Reasoning and factuality
| Benchmark | Metric | Gemma 3 PT 1B | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
| ------------------------------ |----------------|:--------------:|:-------------:|:--------------:|:--------------:|
| [HellaSwag][hellaswag] | 10-shot | 62.3 | 77.2 | 84.2 | 85.6 |
| [BoolQ][boolq] | 0-shot | 63.2 | 72.3 | 78.8 | 82.4 |
| [PIQA][piqa] | 0-shot | 73.8 | 79.6 | 81.8 | 83.3 |
| [SocialIQA][socialiqa] | 0-shot | 48.9 | 51.9 | 53.4 | 54.9 |
| [TriviaQA][triviaqa] | 5-shot | 39.8 | 65.8 | 78.2 | 85.5 |
| [Natural Questions][naturalq] | 5-shot | 9.48 | 20.0 | 31.4 | 36.1 |
| [ARC-c][arc] | 25-shot | 38.4 | 56.2 | 68.9 | 70.6 |
| [ARC-e][arc] | 0-shot | 73.0 | 82.4 | 88.3 | 89.0 |
| [WinoGrande][winogrande] | 5-shot | 58.2 | 64.7 | 74.3 | 78.8 |
| [BIG-Bench Hard][bbh] | few-shot | 28.4 | 50.9 | 72.6 | 77.7 |
| [DROP][drop] | 1-shot | 42.4 | 60.1 | 72.2 | 77.2 |
[hellaswag]: https://arxiv.org/abs/1905.07830
[boolq]: https://arxiv.org/abs/1905.10044
[piqa]: https://arxiv.org/abs/1911.11641
[socialiqa]: https://arxiv.org/abs/1904.09728
[triviaqa]: https://arxiv.org/abs/1705.03551
[naturalq]: https://github.com/google-research-datasets/natural-questions
[arc]: https://arxiv.org/abs/1911.01547
[winogrande]: https://arxiv.org/abs/1907.10641
[bbh]: https://paperswithcode.com/dataset/bbh
[drop]: https://arxiv.org/abs/1903.00161
#### STEM and code
| Benchmark | Metric | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
| ------------------------------ |----------------|:-------------:|:--------------:|:--------------:|
| [MMLU][mmlu] | 5-shot | 59.6 | 74.5 | 78.6 |
| [MMLU][mmlu] (Pro COT) | 5-shot | 29.2 | 45.3 | 52.2 |
| [AGIEval][agieval] | 3-5-shot | 42.1 | 57.4 | 66.2 |
| [MATH][math] | 4-shot | 24.2 | 43.3 | 50.0 |
| [GSM8K][gsm8k] | 8-shot | 38.4 | 71.0 | 82.6 |
| [GPQA][gpqa] | 5-shot | 15.0 | 25.4 | 24.3 |
| [MBPP][mbpp] | 3-shot | 46.0 | 60.4 | 65.6 |
| [HumanEval][humaneval] | 0-shot | 36.0 | 45.7 | 48.8 |
[mmlu]: https://arxiv.org/abs/2009.03300
[agieval]: https://arxiv.org/abs/2304.06364
[math]: https://arxiv.org/abs/2103.03874
[gsm8k]: https://arxiv.org/abs/2110.14168
[gpqa]: https://arxiv.org/abs/2311.12022
[mbpp]: https://arxiv.org/abs/2108.07732
[humaneval]: https://arxiv.org/abs/2107.03374
#### Multilingual
| Benchmark | Gemma 3 PT 1B | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
| ------------------------------------ |:-------------:|:-------------:|:--------------:|:--------------:|
| [MGSM][mgsm] | 2.04 | 34.7 | 64.3 | 74.3 |
| [Global-MMLU-Lite][global-mmlu-lite] | 24.9 | 57.0 | 69.4 | 75.7 |
| [WMT24++][wmt24pp] (ChrF) | 36.7 | 48.4 | 53.9 | 55.7 |
| [FloRes][flores] | 29.5 | 39.2 | 46.0 | 48.8 |
| [XQuAD][xquad] (all) | 43.9 | 68.0 | 74.5 | 76.8 |
| [ECLeKTic][eclektic] | 4.69 | 11.0 | 17.2 | 24.4 |
| [IndicGenBench][indicgenbench] | 41.4 | 57.2 | 61.7 | 63.4 |
[mgsm]: https://arxiv.org/abs/2210.03057
[flores]: https://arxiv.org/abs/2106.03193
[xquad]: https://arxiv.org/abs/1910.11856v3
[global-mmlu-lite]: https://huggingface.co/datasets/CohereForAI/Global-MMLU-Lite
[wmt24pp]: https://arxiv.org/abs/2502.12404v1
[eclektic]: https://arxiv.org/abs/2502.21228
[indicgenbench]: https://arxiv.org/abs/2404.16816
#### Multimodal
| Benchmark | Gemma 3 PT 4B | Gemma 3 PT 12B | Gemma 3 PT 27B |
| ------------------------------ |:-------------:|:--------------:|:--------------:|
| [COCOcap][coco-cap] | 102 | 111 | 116 |
| [DocVQA][docvqa] (val) | 72.8 | 82.3 | 85.6 |
| [InfoVQA][info-vqa] (val) | 44.1 | 54.8 | 59.4 |
| [MMMU][mmmu] (pt) | 39.2 | 50.3 | 56.1 |
| [TextVQA][textvqa] (val) | 58.9 | 66.5 | 68.6 |
| [RealWorldQA][realworldqa] | 45.5 | 52.2 | 53.9 |
| [ReMI][remi] | 27.3 | 38.5 | 44.8 |
| [AI2D][ai2d] | 63.2 | 75.2 | 79.0 |
| [ChartQA][chartqa] | 63.6 | 74.7 | 76.3 |
| [VQAv2][vqav2] | 63.9 | 71.2 | 72.9 |
| [BLINK][blinkvqa] | 38.0 | 35.9 | 39.6 |
| [OKVQA][okvqa] | 51.0 | 58.7 | 60.2 |
| [TallyQA][tallyqa] | 42.5 | 51.8 | 54.3 |
| [SpatialSense VQA][ss-vqa] | 50.9 | 60.0 | 59.4 |
| [CountBenchQA][countbenchqa] | 26.1 | 17.8 | 68.0 |
[coco-cap]: https://cocodataset.org/#home
[docvqa]: https://www.docvqa.org/
[info-vqa]: https://arxiv.org/abs/2104.12756
[mmmu]: https://arxiv.org/abs/2311.16502
[textvqa]: https://textvqa.org/
[realworldqa]: https://paperswithcode.com/dataset/realworldqa
[remi]: https://arxiv.org/html/2406.09175v1
[ai2d]: https://allenai.org/data/diagrams
[chartqa]: https://arxiv.org/abs/2203.10244
[vqav2]: https://visualqa.org/index.html
[blinkvqa]: https://arxiv.org/abs/2404.12390
[okvqa]: https://okvqa.allenai.org/
[tallyqa]: https://arxiv.org/abs/1810.12440
[ss-vqa]: https://arxiv.org/abs/1908.02660
[countbenchqa]: https://github.com/google-research/big_vision/blob/main/big_vision/datasets/countbenchqa/
## Ethics and Safety
Ethics and safety evaluation approach and results.
### Evaluation Approach
Our evaluation methods include structured evaluations and internal red-teaming
testing of relevant content policies. Red-teaming was conducted by a number of
different teams, each with different goals and human evaluation metrics. These
models were evaluated against a number of different categories relevant to
ethics and safety, including:
- **Child Safety**: Evaluation of text-to-text and image to text prompts
covering child safety policies, including child sexual abuse and
exploitation.
- **Content Safety:** Evaluation of text-to-text and image to text prompts
covering safety policies including, harassment, violence and gore, and hate
speech.
- **Representational Harms**: Evaluation of text-to-text and image to text
prompts covering safety policies including bias, stereotyping, and harmful
associations or inaccuracies.
In addition to development level evaluations, we conduct "assurance
evaluations" which are our 'arms-length' internal evaluations for responsibility
governance decision making. They are conducted separately from the model
development team, to inform decision making about release. High level findings
are fed back to the model team, but prompt sets are held-out to prevent
overfitting and preserve the results' ability to inform decision making.
Assurance evaluation results are reported to our Responsibility & Safety Council
as part of release review.
### Evaluation Results
For all areas of safety testing, we saw major improvements in the categories of
child safety, content safety, and representational harms relative to previous
Gemma models. All testing was conducted without safety filters to evaluate the
model capabilities and behaviors. For both text-to-text and image-to-text, and
across all model sizes, the model produced minimal policy violations, and showed
significant improvements over previous Gemma models' performance with respect
to ungrounded inferences. A limitation of our evaluations was they included only
English language prompts.
## Usage and Limitations
These models have certain limitations that users should be aware of.
### Intended Usage
Open vision-language models (VLMs) models have a wide range of applications
across various industries and domains. The following list of potential uses is
not comprehensive. The purpose of this list is to provide contextual information
about the possible use-cases that the model creators considered as part of model
training and development.
- Content Creation and Communication
- Text Generation: These models can be used to generate creative text
formats such as poems, scripts, code, marketing copy, and email drafts.
- Chatbots and Conversational AI: Power conversational interfaces
for customer service, virtual assistants, or interactive applications.
- Text Summarization: Generate concise summaries of a text corpus,
research papers, or reports.
- Image Data Extraction: These models can be used to extract,
interpret, and summarize visual data for text communications.
- Research and Education
- Natural Language Processing (NLP) and VLM Research: These
models can serve as a foundation for researchers to experiment with VLM
and NLP techniques, develop algorithms, and contribute to the
advancement of the field.
- Language Learning Tools: Support interactive language learning
experiences, aiding in grammar correction or providing writing practice.
- Knowledge Exploration: Assist researchers in exploring large
bodies of text by generating summaries or answering questions about
specific topics.
### Limitations
- Training Data
- The quality and diversity of the training data significantly
influence the model's capabilities. Biases or gaps in the training data
can lead to limitations in the model's responses.
- The scope of the training dataset determines the subject areas
the model can handle effectively.
- Context and Task Complexity
- Models are better at tasks that can be framed with clear
prompts and instructions. Open-ended or highly complex tasks might be
challenging.
- A model's performance can be influenced by the amount of context
provided (longer context generally leads to better outputs, up to a
certain point).
- Language Ambiguity and Nuance
- Natural language is inherently complex. Models might struggle
to grasp subtle nuances, sarcasm, or figurative language.
- Factual Accuracy
- Models generate responses based on information they learned
from their training datasets, but they are not knowledge bases. They
may generate incorrect or outdated factual statements.
- Common Sense
- Models rely on statistical patterns in language. They might
lack the ability to apply common sense reasoning in certain situations.
### Ethical Considerations and Risks
The development of vision-language models (VLMs) raises several ethical
concerns. In creating an open model, we have carefully considered the following:
- Bias and Fairness
- VLMs trained on large-scale, real-world text and image data can
reflect socio-cultural biases embedded in the training material. These
models underwent careful scrutiny, input data pre-processing described
and posterior evaluations reported in this card.
- Misinformation and Misuse
- VLMs can be misused to generate text that is false, misleading,
or harmful.
- Guidelines are provided for responsible use with the model, see the
[Responsible Generative AI Toolkit][rai-toolkit].
- Transparency and Accountability:
- This model card summarizes details on the models' architecture,
capabilities, limitations, and evaluation processes.
- A responsibly developed open model offers the opportunity to
share innovation by making VLM technology accessible to developers and
researchers across the AI ecosystem.
Risks identified and mitigations:
- **Perpetuation of biases**: It's encouraged to perform continuous
monitoring (using evaluation metrics, human review) and the exploration of
de-biasing techniques during model training, fine-tuning, and other use
cases.
- **Generation of harmful content**: Mechanisms and guidelines for content
safety are essential. Developers are encouraged to exercise caution and
implement appropriate content safety safeguards based on their specific
product policies and application use cases.
- **Misuse for malicious purposes**: Technical limitations and developer
and end-user education can help mitigate against malicious applications of
VLMs. Educational resources and reporting mechanisms for users to flag
misuse are provided. Prohibited uses of Gemma models are outlined in the
[Gemma Prohibited Use Policy][prohibited-use].
- **Privacy violations**: Models were trained on data filtered for removal
of certain personal information and other sensitive data. Developers are
encouraged to adhere to privacy regulations with privacy-preserving
techniques.
### Benefits
At the time of release, this family of models provides high-performance open
vision-language model implementations designed from the ground up for
responsible AI development compared to similarly sized models.
Using the benchmark evaluation metrics described in this document, these models
have shown to provide superior performance to other, comparably-sized open model
alternatives.
[g3-tech-report]: https://goo.gle/Gemma3Report
[rai-toolkit]: https://ai.google.dev/responsible
[kaggle-gemma]: https://www.kaggle.com/models/google/gemma-3
[vertex-mg-gemma3]: https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma3
[terms]: https://ai.google.dev/gemma/terms
[safety-policies]: https://ai.google/static/documents/ai-responsibility-update-published-february-2025.pdf
[prohibited-use]: https://ai.google.dev/gemma/prohibited_use_policy
[tpu]: https://cloud.google.com/tpu/docs/intro-to-tpu
[sustainability]: https://sustainability.google/operating-sustainably/
[jax]: https://github.com/jax-ml/jax
[ml-pathways]: https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture/
[sustainability]: https://sustainability.google/operating-sustainably/
[gemini-2-paper]: https://arxiv.org/abs/2312.11805 | [
"QUESTION_ANSWERING",
"SUMMARIZATION"
] | Non_BioNLP |
caldana/distilbert-base-uncased-finetuned-emotion | caldana | text-classification | [
"transformers",
"pytorch",
"tensorboard",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,654,035,418,000 | 2022-05-31T23:07:12 | 10 | 0 | ---
datasets:
- emotion
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-emotion
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: emotion
type: emotion
args: default
metrics:
- type: accuracy
value: 0.927
name: Accuracy
- type: f1
value: 0.927055679622598
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-emotion
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2236
- Accuracy: 0.927
- F1: 0.9271
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.8251 | 1.0 | 250 | 0.3264 | 0.9015 | 0.8981 |
| 0.2534 | 2.0 | 500 | 0.2236 | 0.927 | 0.9271 |
### Framework versions
- Transformers 4.11.3
- Pytorch 1.11.0
- Datasets 1.16.1
- Tokenizers 0.10.3
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
RichardErkhov/NbAiLab_-_nb-llama-3.2-1B-awq | RichardErkhov | null | [
"safetensors",
"llama",
"4-bit",
"awq",
"region:us"
] | 1,736,746,074,000 | 2025-01-13T05:28:31 | 5 | 0 | ---
{}
---
Quantization made by Richard Erkhov.
[Github](https://github.com/RichardErkhov)
[Discord](https://discord.gg/pvy7H8DZMG)
[Request more models](https://github.com/RichardErkhov/quant_request)
nb-llama-3.2-1B - AWQ
- Model creator: https://huggingface.co/NbAiLab/
- Original model: https://huggingface.co/NbAiLab/nb-llama-3.2-1B/
Original model description:
---
language:
- no # Generic Norwegian
- nb # Norwegian Bokmål
- nn # Norwegian Nynorsk
- en # English
- sv # Swedish
- da # Danish
tags:
- norwegian
- bokmål
- nynorsk
- swedish
- danish
- multilingual
- text-generation
pipeline_tag: text-generation
license: llama3.2
---
## Model Card: NB-Llama-3.2-1B
---
### Model Overview
**NB-Llama-3.2-1B** is part of the **NB-Llama-3.2** series of models, trained on top of [Llama-3.2-1B](https://huggingface.co/meta-llama/Llama-3.2-1B). This multilingual generative model was fine-tuned specifically to support Norwegian Bokmål, Norwegian Nynorsk, and English, with partial support for Swedish and Danish.
The basic idea with this model series was to explore how current state-of-the-art models could be improved for Norwegian by training only on publicly available data. While these models are trained by the National Library of Norway, they do not include data only available through legal deposit. They do, however, contain public data like governmental reports that are both publicly available and legally deposited.
---
### Key Features
- **Base Model**: Built on Llama-3.2-1B.
- **Languages**:
- Full support: Norwegian Bokmål (nb), Norwegian Nynorsk (nn), English (en).
- Partial support: Swedish (sv), Danish (da).
- **Purpose**: Supports Norwegian-specific tasks such as question-answering, summarization, and language modeling, while being capable of multilingual generation and translation. Efforts have been made to preserve the English capabilities from the underlying Meta Llama model.
- **Training Data**: Combines publicly available multilingual datasets with synthetic data generation, focusing on Norwegian, English, Swedish, and Danish sources. Additional details are provided below.
- **Architecture**: The model uses the Llama 3.2 architecture. It is an auto-regressive language model with an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and Direct Preference Optimization (DPO) for alignment.
---
### Model Details
- **Developer**: National Library of Norway (NB-AiLab).
- **Parameters**: 1 billion.
- **Knowledge Cutoff**: May 2024.
- **License**: [Llama 3.2 Community License](https://github.com/meta-llama/llama-models/blob/main/models/llama3.2/LICENSE).
---
### Motivation
The primary goal of **NB-Llama-3.2-1B** is to advance support for Norwegian language technologies and strengthen support for Norwegian Bokmål and Norwegian Nynorsk. Since much knowledge and culture are also expressed in English, Swedish, and Danish, open sources in these languages are included in the training datasets when possible.
---
### Intended Use
#### Use Cases
- Dialogue systems.
- General multilingual text generation and language modeling.
- Norwegian-specific tasks such as:
- Summarization of texts in Bokmål or Nynorsk.
- Question-answering tailored to Norwegian cultural and linguistic contexts.
#### Out-of-Scope
- Use in violation of applicable laws or regulations.
- Tasks outside the supported languages without additional fine-tuning.
- High-risk domains without appropriate safety measures.
---
### How to Use
Please note tht this is still a research project, and the purpose of releasing the models are to investigate the potential in adapting these models for Norwegian language. The intented use case is experiemental. For end-users, we strongly recommend using the instruction-tuned models. We provide quantized models with close to the same accuracy that will run much faster on most platforms. When fine-tuning the instruction-tuned models, best results are obtained when applying the appropriate templates from Llama 3.2.
#### Using `transformers`
```python
import transformers
model_id = "NbAiLab/nb-llama-3.2-1B"
pipeline = transformers.pipeline(
"text-generation",
model=model_id,
model_kwargs={"torch_dtype": "bfloat16"},
device_map="auto"
)
output = pipeline("Hva er Nasjonalbibliotekets rolle i AI-utvikling?")
print(output)
```
---
### Training Data
**Overview:**
The training data is based entirely on publicly available datasets and synthetically generated data. A key aspect of the training process was leveraging high-quality knowledge sources in Norwegian, English, Swedish, and Danish.
Parts of the following publicly available datasets were used:
- [CulturaX](https://huggingface.co/datasets/uonlp/CulturaX)
- [High Performance Language Technologies (HPLT)](https://huggingface.co/datasets/HPLT/hplt_monolingual_v1_2)
- [Norwegian Colossal Corpus (NCC)](https://huggingface.co/datasets/NCC/Norwegian-Colossal-Corpus)
- [Wikipedia](https://huggingface.co/datasets/wikimedia/wikipedia)
---
### Data Selection
To ensure the highest quality training data, only a small subset of the original raw data was used. [Corpus Quality Classifiers](https://huggingface.co/collections/NbAiLab/corpus-quality-classifier-673f15926c2774fcc88f23aa) built on [nb-bert-base](https://huggingface.co/NbAiLab/nb-bert-base) were trained to evaluate both educational value and linguistic quality of the training samples. These models are released along with the NB-Llama-3.x models, and are considered the main output from this initiative.
- **Categorization Methods:**
- Inspired by the [FineWeb](https://example.com/FineWeb) project.
- Evaluated for:
- **Educational Value:** Prioritizing high-value training samples.
- **Linguistic Quality:** Ensuring clarity and accuracy in training data.
- **Guidance and Release:**
- Categorization was guided by insights from [Gemini 1.5](https://blog.google/technology/ai/google-gemini-next-generation-model-february-2024/#gemini-15).
- The classifiers are released alongside this model and are [available here](https://classifier-release-link-here).
---
### Licensing
The model is released under the [Llama 3.2 Community License](https://github.com/meta-llama/llama-models/blob/main/models/llama3.2/LICENSE), allowing for research and commercial use within defined limitations. Refer to the [Acceptable Use Policy](https://llama.meta.com/llama3.2/use-policy) for specific restrictions.
---
### Citing & Authors
The model was trained and documentation written by Per Egil Kummervold as part of the NoTraM-project.
### Funding and Acknowledgement
Training this model was supported by Google’s TPU Research Cloud (TRC), which generously supplied us with Cloud TPUs essential for our computational
needs.
| [
"TRANSLATION",
"SUMMARIZATION"
] | Non_BioNLP |
chriswilson2020/distilbert-base-uncased-finetuned-emotion | chriswilson2020 | text-classification | [
"transformers",
"safetensors",
"distilbert",
"text-classification",
"generated_from_trainer",
"dataset:emotion",
"base_model:distilbert/distilbert-base-uncased",
"base_model:finetune:distilbert/distilbert-base-uncased",
"license:apache-2.0",
"model-index",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,713,623,479,000 | 2024-04-20T14:52:36 | 4 | 0 | ---
base_model: distilbert-base-uncased
datasets:
- emotion
license: apache-2.0
metrics:
- accuracy
- f1
tags:
- generated_from_trainer
model-index:
- name: distilbert-base-uncased-finetuned-emotion
results:
- task:
type: text-classification
name: Text Classification
dataset:
name: emotion
type: emotion
config: split
split: validation
args: split
metrics:
- type: accuracy
value: 0.9225
name: Accuracy
- type: f1
value: 0.9224892356710013
name: F1
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# distilbert-base-uncased-finetuned-emotion
This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the emotion dataset.
It achieves the following results on the evaluation set:
- Loss: 0.2128
- Accuracy: 0.9225
- F1: 0.9225
## Model description
More information needed
## Intended uses & limitations
More information needed
## Training and evaluation data
More information needed
## Training procedure
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 2
### Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:------:|
| 0.8397 | 1.0 | 250 | 0.3091 | 0.91 | 0.9098 |
| 0.243 | 2.0 | 500 | 0.2128 | 0.9225 | 0.9225 |
### Framework versions
- Transformers 4.40.0
- Pytorch 2.2.2+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1
| [
"TEXT_CLASSIFICATION"
] | Non_BioNLP |
eligapris/kin-eng | eligapris | translation | [
"transformers",
"pytorch",
"m2m_100",
"text2text-generation",
"translation",
"en",
"rw",
"dataset:mbazaNLP/NMT_Tourism_parallel_data_en_kin",
"dataset:mbazaNLP/NMT_Education_parallel_data_en_kin",
"dataset:mbazaNLP/Kinyarwanda_English_parallel_dataset",
"license:cc-by-2.0",
"autotrain_compatible",
"endpoints_compatible",
"region:us"
] | 1,725,482,595,000 | 2024-09-05T00:00:07 | 0 | 0 | ---
datasets:
- mbazaNLP/NMT_Tourism_parallel_data_en_kin
- mbazaNLP/NMT_Education_parallel_data_en_kin
- mbazaNLP/Kinyarwanda_English_parallel_dataset
language:
- en
- rw
library_name: transformers
license: cc-by-2.0
pipeline_tag: translation
---
## Model Details
### Model Description
<!-- Provide a longer summary of what this model is. -->
This is a Machine Translation model, finetuned from [NLLB](https://huggingface.co/facebook/nllb-200-distilled-1.3B)-200's distilled 1.3B model, it is meant to be used in machine translation for education-related data.
- **Finetuning code repository:** the code used to finetune this model can be found [here](https://github.com/Digital-Umuganda/twb_nllb_finetuning)
<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
## How to Get Started with the Model
Use the code below to get started with the model.
### Training Procedure
The model was finetuned on three datasets; a [general](https://huggingface.co/datasets/mbazaNLP/Kinyarwanda_English_parallel_dataset) purpose dataset, a [tourism](https://huggingface.co/datasets/mbazaNLP/NMT_Tourism_parallel_data_en_kin), and an [education](https://huggingface.co/datasets/mbazaNLP/NMT_Education_parallel_data_en_kin) dataset.
The model was finetuned in two phases.
#### Phase one:
- General purpose dataset
- Education dataset
- Tourism dataset
#### Phase two:
- Education dataset
Other than the dataset changes between phase one, and phase two finetuning; no other hyperparameters were modified. In both cases, the model was trained on an A100 40GB GPU for two epochs.
## Evaluation
<!-- This section describes the evaluation protocols and provides the results. -->
<!-- This should link to a Data Card if possible. -->
#### Metrics
Model performance was measured using BLEU, spBLEU, TER, and chrF++ metrics.
### Results
|Lang. Direction| BLEU | spBLEU | chrf++ |TER |
|:----|:----:|:----:|:----:|----:|
| Eng -> Kin | 45.96 | 59.20 | 68.79 | 41.61 |
| Kin -> Eng | 43.98 | 44.94 | 63.05 | 41.41 |
<!-- [More Information Needed] --> | [
"TRANSLATION"
] | Non_BioNLP |
Subsets and Splits