Kaleidoscope: In-language Exams for Massively Multilingual Vision Evaluation
Paper
•
2504.07072
•
Published
•
9
NLP en ES
non-profit
AI & ML interests
Ahora la comunidad se llama Somos NLP ➡️ https://huggingface.co/somosnlp
Recent Activity
View all activity
nlp-en-es's activity
mariagrandury
authored
2
papers
25 days ago
lewtun
authored
a
paper
about 1 month ago
lewtun
authored
a
paper
2 months ago
osanseviero
authored
a
paper
3 months ago
lewtun
authored
a
paper
3 months ago
Post
2807
Introducing OlympicCoder: a series of open reasoning models that can solve olympiad-level programming problems 🧑💻
- 7B open-r1/OlympicCoder-7B
- 32B open-r1/OlympicCoder-32B
We find that OlympicCoder models outperform Claude 3.7 Sonnet, as well as others over 100x larger 💪
Together with the models, we are releasing:
📊CodeForces-CoTs: new dataset of code problems from the most popular competitive coding platform, with R1 traces in C++ and Python open-r1/codeforces-cots
🏆 IOI'2024: a new benchmark of VERY hard programming problems where even frontier models struggle to match human performance open-r1/ioi
For links to the models and datasets, check out our latest progress report from Open R1: https://huggingface.co/blog/open-r1/update-3
- 7B open-r1/OlympicCoder-7B
- 32B open-r1/OlympicCoder-32B
We find that OlympicCoder models outperform Claude 3.7 Sonnet, as well as others over 100x larger 💪
Together with the models, we are releasing:
📊CodeForces-CoTs: new dataset of code problems from the most popular competitive coding platform, with R1 traces in C++ and Python open-r1/codeforces-cots
🏆 IOI'2024: a new benchmark of VERY hard programming problems where even frontier models struggle to match human performance open-r1/ioi
For links to the models and datasets, check out our latest progress report from Open R1: https://huggingface.co/blog/open-r1/update-3
- 1 reply
Post
5289
Introducing OpenR1-Math-220k!
open-r1/OpenR1-Math-220k
The community has been busy distilling DeepSeek-R1 from inference providers, but we decided to have a go at doing it ourselves from scratch 💪
What’s new compared to existing reasoning datasets?
♾ Based on AI-MO/NuminaMath-1.5: we focus on math reasoning traces and generate answers for problems in NuminaMath 1.5, an improved version of the popular NuminaMath-CoT dataset.
🐳 800k R1 reasoning traces: We generate two answers for 400k problems using DeepSeek R1. The filtered dataset contains 220k problems with correct reasoning traces.
📀 512 H100s running locally: Instead of relying on an API, we leverage vLLM and SGLang to run generations locally on our science cluster, generating 180k reasoning traces per day.
⏳ Automated filtering: We apply Math Verify to only retain problems with at least one correct answer. We also leverage Llama3.3-70B-Instruct as a judge to retrieve more correct examples (e.g for cases with malformed answers that can’t be verified with a rules-based parser)
📊 We match the performance of DeepSeek-Distill-Qwen-7B by finetuning Qwen-7B-Math-Instruct on our dataset.
🔎 Read our blog post for all the nitty gritty details: https://huggingface.co/blog/open-r1/update-2
open-r1/OpenR1-Math-220k
The community has been busy distilling DeepSeek-R1 from inference providers, but we decided to have a go at doing it ourselves from scratch 💪
What’s new compared to existing reasoning datasets?
♾ Based on AI-MO/NuminaMath-1.5: we focus on math reasoning traces and generate answers for problems in NuminaMath 1.5, an improved version of the popular NuminaMath-CoT dataset.
🐳 800k R1 reasoning traces: We generate two answers for 400k problems using DeepSeek R1. The filtered dataset contains 220k problems with correct reasoning traces.
📀 512 H100s running locally: Instead of relying on an API, we leverage vLLM and SGLang to run generations locally on our science cluster, generating 180k reasoning traces per day.
⏳ Automated filtering: We apply Math Verify to only retain problems with at least one correct answer. We also leverage Llama3.3-70B-Instruct as a judge to retrieve more correct examples (e.g for cases with malformed answers that can’t be verified with a rules-based parser)
📊 We match the performance of DeepSeek-Distill-Qwen-7B by finetuning Qwen-7B-Math-Instruct on our dataset.
🔎 Read our blog post for all the nitty gritty details: https://huggingface.co/blog/open-r1/update-2
lewtun
authored
a
paper
4 months ago
Post
10404
We are reproducing the full DeepSeek R1 data and training pipeline so everybody can use their recipe. Instead of doing it in secret we can do it together in the open!
🧪 Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1.
🧠 Step 2: replicate the pure RL pipeline that DeepSeek used to create R1-Zero. This will involve curating new, large-scale datasets for math, reasoning, and code.
🔥 Step 3: show we can go from base model -> SFT -> RL via multi-stage training.
Follow along: https://github.com/huggingface/open-r1
🧪 Step 1: replicate the R1-Distill models by distilling a high-quality reasoning corpus from DeepSeek-R1.
🧠 Step 2: replicate the pure RL pipeline that DeepSeek used to create R1-Zero. This will involve curating new, large-scale datasets for math, reasoning, and code.
🔥 Step 3: show we can go from base model -> SFT -> RL via multi-stage training.
Follow along: https://github.com/huggingface/open-r1
- 5 replies
Post
3961
I was initially pretty sceptical about Meta's Coconut paper [1] because the largest perf gains were reported on toy linguistic problems. However, these results on machine translation are pretty impressive!
https://x.com/casper_hansen_/status/1875872309996855343
Together with the recent PRIME method [2] for scaling RL, reasoning for open models is looking pretty exciting for 2025!
[1] Training Large Language Models to Reason in a Continuous Latent Space (2412.06769)
[2] https://huggingface.co/blog/ganqu/prime
https://x.com/casper_hansen_/status/1875872309996855343
Together with the recent PRIME method [2] for scaling RL, reasoning for open models is looking pretty exciting for 2025!
[1] Training Large Language Models to Reason in a Continuous Latent Space (2412.06769)
[2] https://huggingface.co/blog/ganqu/prime
Post
2350
This paper (
HuatuoGPT-o1, Towards Medical Complex Reasoning with LLMs (2412.18925)) has a really interesting recipe for inducing o1-like behaviour in Llama models:
* Iteratively sample CoTs from the model, using a mix of different search strategies. This gives you something like Stream of Search via prompting.
* Verify correctness of each CoT using GPT-4o (needed because exact match doesn't work well in medicine where there are lots of aliases)
* Use GPT-4o to reformat the concatenated CoTs into a single stream that includes smooth transitions like "hmm, wait" etc that one sees in o1
* Use the resulting data for SFT & RL
* Use sparse rewards from GPT-4o to guide RL training. They find RL gives an average ~3 point boost across medical benchmarks and SFT on this data already gives a strong improvement.
Applying this strategy to other domains could be quite promising, provided the training data can be formulated with verifiable problems!
* Iteratively sample CoTs from the model, using a mix of different search strategies. This gives you something like Stream of Search via prompting.
* Verify correctness of each CoT using GPT-4o (needed because exact match doesn't work well in medicine where there are lots of aliases)
* Use GPT-4o to reformat the concatenated CoTs into a single stream that includes smooth transitions like "hmm, wait" etc that one sees in o1
* Use the resulting data for SFT & RL
* Use sparse rewards from GPT-4o to guide RL training. They find RL gives an average ~3 point boost across medical benchmarks and SFT on this data already gives a strong improvement.
Applying this strategy to other domains could be quite promising, provided the training data can be formulated with verifiable problems!
- 1 reply
Post
7008
We outperform Llama 70B with Llama 3B on hard math by scaling test-time compute 🔥
How? By combining step-wise reward models with tree search algorithms :)
We show that smol models can match or exceed the performance of their much larger siblings when given enough "time to think"
We're open sourcing the full recipe and sharing a detailed blog post.
In our blog post we cover:
📈 Compute-optimal scaling: How we implemented DeepMind's recipe to boost the mathematical capabilities of open models at test-time.
🎄 Diverse Verifier Tree Search (DVTS): An unpublished extension we developed to the verifier-guided tree search technique. This simple yet effective method improves diversity and delivers better performance, particularly at large test-time compute budgets.
🧭 Search and Learn: A lightweight toolkit for implementing search strategies with LLMs and built for speed with vLLM
Here's the links:
- Blog post: HuggingFaceH4/blogpost-scaling-test-time-compute
- Code: https://github.com/huggingface/search-and-learn
Enjoy!
How? By combining step-wise reward models with tree search algorithms :)
We show that smol models can match or exceed the performance of their much larger siblings when given enough "time to think"
We're open sourcing the full recipe and sharing a detailed blog post.
In our blog post we cover:
📈 Compute-optimal scaling: How we implemented DeepMind's recipe to boost the mathematical capabilities of open models at test-time.
🎄 Diverse Verifier Tree Search (DVTS): An unpublished extension we developed to the verifier-guided tree search technique. This simple yet effective method improves diversity and delivers better performance, particularly at large test-time compute budgets.
🧭 Search and Learn: A lightweight toolkit for implementing search strategies with LLMs and built for speed with vLLM
Here's the links:
- Blog post: HuggingFaceH4/blogpost-scaling-test-time-compute
- Code: https://github.com/huggingface/search-and-learn
Enjoy!
- 2 replies
mariagrandury
authored
a
paper
9 months ago
mariagrandury
authored
a
paper
11 months ago
Post
6447
🚨Exciting news for the Multilingual Synthetic Data Community!🚨
I’ve taken inspiration from the MAGPIE paper on Llama-3-8B-instruct and extended its capabilities. Here’s what’s new!
🗞 The MAGPIE paper showcased that if you use the instruction-tuned version (
🤔 While reading a script by Sebastian Raschka, PhD, I wondered: Could these advancements be replicated in other languages? Specifically, could they benefit non-English datasets?
🎉 And the answer is YES! At least for Spanish. I've successfully adapted the techniques for Spanish, proving the model's flexibility and multilingual capabilities.
👩💻 To make this accessible, I created a basic script (heavily inspired by the Sebastian Raschka one) that allows you to generate similar datasets using
[Script](https://gist.github.com/mrm8488/4650a5e3cc45523798a527a3446eb312)
🔍 Explore the datasets 📚 generated using our new script!
- [Llama-3-8B](https://huggingface.co/datasets/mrm8488/dataset_llama3_5000_samples_es_4231_filtered)
- [Phi-3-medium](https://huggingface.co/datasets/mrm8488/dataset_phi3-medium_5000_samples_es_3906_filtered)
- [Phi-3-mini](https://huggingface.co/datasets/mrm8488/dataset_phi3_5000_samples_es_3282_filtered)
Note: These datasets have basic filtering. Apply additional quality filters before using them to fine-tune large language models.
Inspiration and base script:
https://github.com/rasbt/LLMs-from-scratch/blob/main/ch07/05_dataset-generation/llama3-ollama.ipynb
https://www.linkedin.com/feed/update/urn:li:activity:7210982019751661568/
I’ve taken inspiration from the MAGPIE paper on Llama-3-8B-instruct and extended its capabilities. Here’s what’s new!
🗞 The MAGPIE paper showcased that if you use the instruction-tuned version (
Llama-3-8B-instruct) to generate synthetic instructions and then fine-tune the base version (Llama-3-8B) on this dataset, you can improve even the it-tuned version🤔 While reading a script by Sebastian Raschka, PhD, I wondered: Could these advancements be replicated in other languages? Specifically, could they benefit non-English datasets?
🎉 And the answer is YES! At least for Spanish. I've successfully adapted the techniques for Spanish, proving the model's flexibility and multilingual capabilities.
👩💻 To make this accessible, I created a basic script (heavily inspired by the Sebastian Raschka one) that allows you to generate similar datasets using
ollama models (initially phi and llama3) automatically and upload it to the Hugging Face Hub
🔍 Explore the datasets 📚 generated using our new script!
- [Llama-3-8B](https://huggingface.co/datasets/mrm8488/dataset_llama3_5000_samples_es_4231_filtered)
- [Phi-3-medium](https://huggingface.co/datasets/mrm8488/dataset_phi3-medium_5000_samples_es_3906_filtered)
- [Phi-3-mini](https://huggingface.co/datasets/mrm8488/dataset_phi3_5000_samples_es_3282_filtered)
Note: These datasets have basic filtering. Apply additional quality filters before using them to fine-tune large language models.
Inspiration and base script:
https://github.com/rasbt/LLMs-from-scratch/blob/main/ch07/05_dataset-generation/llama3-ollama.ipynb
https://www.linkedin.com/feed/update/urn:li:activity:7210982019751661568/
-
- 7 replies
mariagrandury
authored
a
paper
12 months ago
Post
7309
Working on a concept
The ckpt and training code will be soon on the hub.
GPT-2 (small) that uses KANs instead of MLPs.The ckpt and training code will be soon on the hub.
-
- 6 replies
osanseviero
posted
an
update
about 1 year ago
Post
13905
Diaries of Open Source. Part 15 🤗
🕵️♀️Idefics 2 is out, a multimodal open-source model with very nice capabilities
Models, demo, and datasets: HuggingFaceM4/idefics2-661d1971b7c50831dd3ce0fe
Blog: https://hf.co/blog/idefics2
💾Snowflake released snowflake-arctic-embed, a family of powerful small embedding models
Model: Snowflake/snowflake-arctic-embed-m
Blog: https://www.snowflake.com/blog/introducing-snowflake-arctic-embed-snowflakes-state-of-the-art-text-embedding-family-of-models/
✨Pile-T5, EleutherAI's T5 model trained on 2T tokens
Blog: https://blog.eleuther.ai/pile-t5/
Models: EleutherAI/pile-t5-65a76a0d0022dd270b385a66
GitHub: https://github.com/EleutherAI/improved-t5
🤖CodeQwen1.5-7B base and chat models. Models trained on 3T tokens strong benchmark results for code generation, editing and SQL
Blog post: https://qwenlm.github.io/blog/codeqwen1.5/
Demo: https://hf.co/spaces/Qwen/CodeQwen1.5-7b-Chat-demo
Models: Qwen/CodeQwen1.5-7B and Qwen/CodeQwen1.5-7B-Chat
Misc
🦉 DocOwl1.5: Unified Stucture Learning for OCR-free Document Understanding mPLUG/DocOwl
👀Cerule - a tiny Vision LM model Tensoic/Cerule-v0.1
ChemLLM - a LLM for chemistry and molecule science ⚗️https://hf.co/AI4Chem/ChemLLM-7B-Chat-1.5-DPO
Distil Whisper Large
📝New pdf/OCR datasets with 19 samples pixparse/pdf-document-ocr-datasets-660701430b0346f97c4bc628
🔥Gretel AI high quality text-to-sql synthetic dataset gretelai/synthetic_text_to_sql
🕵️♀️Idefics 2 is out, a multimodal open-source model with very nice capabilities
Models, demo, and datasets: HuggingFaceM4/idefics2-661d1971b7c50831dd3ce0fe
Blog: https://hf.co/blog/idefics2
💾Snowflake released snowflake-arctic-embed, a family of powerful small embedding models
Model: Snowflake/snowflake-arctic-embed-m
Blog: https://www.snowflake.com/blog/introducing-snowflake-arctic-embed-snowflakes-state-of-the-art-text-embedding-family-of-models/
✨Pile-T5, EleutherAI's T5 model trained on 2T tokens
Blog: https://blog.eleuther.ai/pile-t5/
Models: EleutherAI/pile-t5-65a76a0d0022dd270b385a66
GitHub: https://github.com/EleutherAI/improved-t5
🤖CodeQwen1.5-7B base and chat models. Models trained on 3T tokens strong benchmark results for code generation, editing and SQL
Blog post: https://qwenlm.github.io/blog/codeqwen1.5/
Demo: https://hf.co/spaces/Qwen/CodeQwen1.5-7b-Chat-demo
Models: Qwen/CodeQwen1.5-7B and Qwen/CodeQwen1.5-7B-Chat
Misc
🦉 DocOwl1.5: Unified Stucture Learning for OCR-free Document Understanding mPLUG/DocOwl
👀Cerule - a tiny Vision LM model Tensoic/Cerule-v0.1
ChemLLM - a LLM for chemistry and molecule science ⚗️https://hf.co/AI4Chem/ChemLLM-7B-Chat-1.5-DPO
Distil Whisper Large
📝New pdf/OCR datasets with 19 samples pixparse/pdf-document-ocr-datasets-660701430b0346f97c4bc628
🔥Gretel AI high quality text-to-sql synthetic dataset gretelai/synthetic_text_to_sql
-
- 5 replies
Post
5108
Introducing Zephyr 141B-A35B 🪁:
HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1
Yesterday, Mistral released their latest base model (via magnet link of course 😅) and the community quickly converted it to transformers format and pushed it to the Hub: mistral-community/Mixtral-8x22B-v0.1
Early evals of this model looked extremely strong, so we teamed up with Argilla and KAIST AI to cook up a Zephyr recipe with a few new alignment techniques that came out recently:
🧑🍳 Align the base model with Odds Ratio Preference Optimisation (ORPO). This novel algorithm developed by @JW17 and @nlee-208 and @j6mes and does not require an SFT step to achieve high performance and is thus much more computationally efficient than methods like DPO and PPO.
🦫 Use a brand new dataset of 7k high-quality, multi-turn preferences that has been developed by our friends at Argilla. To create this dataset, they took the excellent Capybara SFT dataset from @LDJnr LDJnr/Capybara and converted it into a preference dataset by augmenting the final turn with responses from new LLMs that were then ranked by GPT-4.
What we find especially neat about this approach is that training on 7k samples only takes ~1.3h on 4 H100 nodes, yet produces a model that is very strong on chat benchmarks like IFEval and BBH.
Kudos to @alvarobartt @JW17 and @nlee-208 for this very nice and fast-paced collab!
For more details on the paper and dataset, checkout our collection: HuggingFaceH4/zephyr-orpo-6617eba2c5c0e2cc3c151524
HuggingFaceH4/zephyr-orpo-141b-A35b-v0.1
Yesterday, Mistral released their latest base model (via magnet link of course 😅) and the community quickly converted it to transformers format and pushed it to the Hub: mistral-community/Mixtral-8x22B-v0.1
Early evals of this model looked extremely strong, so we teamed up with Argilla and KAIST AI to cook up a Zephyr recipe with a few new alignment techniques that came out recently:
🧑🍳 Align the base model with Odds Ratio Preference Optimisation (ORPO). This novel algorithm developed by @JW17 and @nlee-208 and @j6mes and does not require an SFT step to achieve high performance and is thus much more computationally efficient than methods like DPO and PPO.
🦫 Use a brand new dataset of 7k high-quality, multi-turn preferences that has been developed by our friends at Argilla. To create this dataset, they took the excellent Capybara SFT dataset from @LDJnr LDJnr/Capybara and converted it into a preference dataset by augmenting the final turn with responses from new LLMs that were then ranked by GPT-4.
What we find especially neat about this approach is that training on 7k samples only takes ~1.3h on 4 H100 nodes, yet produces a model that is very strong on chat benchmarks like IFEval and BBH.
Kudos to @alvarobartt @JW17 and @nlee-208 for this very nice and fast-paced collab!
For more details on the paper and dataset, checkout our collection: HuggingFaceH4/zephyr-orpo-6617eba2c5c0e2cc3c151524