File size: 7,859 Bytes
9d76f33 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 |
---
license: apache-2.0
base_model: allenai/longformer-base-4096
tags:
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: longformer-one-step
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. -->
# longformer-one-step
This model is a fine-tuned version of [allenai/longformer-base-4096](https://huggingface.co/allenai/longformer-base-4096) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.4163
- B-claim: {'precision': 0.5555555555555556, 'recall': 0.551948051948052, 'f1-score': 0.5537459283387622, 'support': 154.0}
- B-majorclaim: {'precision': 0.6285714285714286, 'recall': 0.6875, 'f1-score': 0.6567164179104478, 'support': 64.0}
- B-premise: {'precision': 0.7458677685950413, 'recall': 0.8414918414918415, 'f1-score': 0.7907995618838992, 'support': 429.0}
- I-claim: {'precision': 0.6674840608141246, 'recall': 0.6067766384306732, 'f1-score': 0.6356842596917328, 'support': 2243.0}
- I-majorclaim: {'precision': 0.7213656387665198, 'recall': 0.7511467889908257, 'f1-score': 0.7359550561797752, 'support': 872.0}
- I-premise: {'precision': 0.8961770096884001, 'recall': 0.9113300492610837, 'f1-score': 0.903690012542082, 'support': 7511.0}
- O: {'precision': 0.9139617607825701, 'recall': 0.910117334514058, 'f1-score': 0.9120354963948973, 'support': 4517.0}
- Accuracy: 0.8526
- Macro avg: {'precision': 0.7327118889676628, 'recall': 0.751472957805219, 'f1-score': 0.741232390420228, 'support': 15790.0}
- Weighted avg: {'precision': 0.8506339314975508, 'recall': 0.8525649145028499, 'f1-score': 0.8512623407634757, 'support': 15790.0}
## 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: 8
- eval_batch_size: 8
- seed: 42
- 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 | B-claim | B-majorclaim | B-premise | I-claim | I-majorclaim | I-premise | O | Accuracy | Macro avg | Weighted avg |
|:-------------:|:-----:|:----:|:---------------:|:-------------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------:|:-----------------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------------------:|:-----------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------------:|:------------------------------------------------------------------------------------------------------------------:|:--------:|:-------------------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------------------:|
| No log | 1.0 | 196 | 0.4925 | {'precision': 0.4489795918367347, 'recall': 0.14285714285714285, 'f1-score': 0.21674876847290642, 'support': 154.0} | {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 64.0} | {'precision': 0.6521739130434783, 'recall': 0.8741258741258742, 'f1-score': 0.7470119521912351, 'support': 429.0} | {'precision': 0.6343843843843844, 'recall': 0.37672759696834596, 'f1-score': 0.4727272727272727, 'support': 2243.0} | {'precision': 0.6209944751381216, 'recall': 0.6444954128440367, 'f1-score': 0.6325267304445695, 'support': 872.0} | {'precision': 0.845458984375, 'recall': 0.92211423245906, 'f1-score': 0.8821244348213716, 'support': 7511.0} | {'precision': 0.8657378087397086, 'recall': 0.9079034757582466, 'f1-score': 0.8863194294359196, 'support': 4517.0} | 0.8126 | {'precision': 0.5811041653596325, 'recall': 0.5526033907161009, 'f1-score': 0.5482083697276107, 'support': 15790.0} | {'precision': 0.7963354614345164, 'recall': 0.8126029132362255, 'f1-score': 0.7976491141364899, 'support': 15790.0} |
| No log | 2.0 | 392 | 0.4278 | {'precision': 0.5454545454545454, 'recall': 0.5064935064935064, 'f1-score': 0.5252525252525252, 'support': 154.0} | {'precision': 0.7090909090909091, 'recall': 0.609375, 'f1-score': 0.6554621848739497, 'support': 64.0} | {'precision': 0.6920222634508348, 'recall': 0.8694638694638694, 'f1-score': 0.7706611570247933, 'support': 429.0} | {'precision': 0.6465288818229995, 'recall': 0.5439144003566652, 'f1-score': 0.5907990314769976, 'support': 2243.0} | {'precision': 0.7232558139534884, 'recall': 0.713302752293578, 'f1-score': 0.7182448036951501, 'support': 872.0} | {'precision': 0.8695761223977928, 'recall': 0.9231793369724404, 'f1-score': 0.8955763642234421, 'support': 7511.0} | {'precision': 0.9275161588180979, 'recall': 0.8895284480850122, 'f1-score': 0.9081252118883489, 'support': 4517.0} | 0.8413 | {'precision': 0.7304920992840954, 'recall': 0.7221796162378674, 'f1-score': 0.7234458969193153, 'support': 15790.0} | {'precision': 0.8377504411405843, 'recall': 0.8412919569347689, 'f1-score': 0.8381000288341659, 'support': 15790.0} |
| 0.598 | 3.0 | 588 | 0.4163 | {'precision': 0.5555555555555556, 'recall': 0.551948051948052, 'f1-score': 0.5537459283387622, 'support': 154.0} | {'precision': 0.6285714285714286, 'recall': 0.6875, 'f1-score': 0.6567164179104478, 'support': 64.0} | {'precision': 0.7458677685950413, 'recall': 0.8414918414918415, 'f1-score': 0.7907995618838992, 'support': 429.0} | {'precision': 0.6674840608141246, 'recall': 0.6067766384306732, 'f1-score': 0.6356842596917328, 'support': 2243.0} | {'precision': 0.7213656387665198, 'recall': 0.7511467889908257, 'f1-score': 0.7359550561797752, 'support': 872.0} | {'precision': 0.8961770096884001, 'recall': 0.9113300492610837, 'f1-score': 0.903690012542082, 'support': 7511.0} | {'precision': 0.9139617607825701, 'recall': 0.910117334514058, 'f1-score': 0.9120354963948973, 'support': 4517.0} | 0.8526 | {'precision': 0.7327118889676628, 'recall': 0.751472957805219, 'f1-score': 0.741232390420228, 'support': 15790.0} | {'precision': 0.8506339314975508, 'recall': 0.8525649145028499, 'f1-score': 0.8512623407634757, 'support': 15790.0} |
### Framework versions
- Transformers 4.33.0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
|