bert-small-finetuned-cuad-longer
This model is a fine-tuned version of google/bert_uncased_L-4_H-512_A-8 on the cuad dataset. It achieves the following results on the evaluation set:
- Loss: 0.0617
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: 10
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.0601 | 1.0 | 8702 | 0.0404 |
| 0.0551 | 2.0 | 17404 | 0.0394 |
| 0.0428 | 3.0 | 26106 | 0.0481 |
| 0.0375 | 4.0 | 34808 | 0.0425 |
| 0.0403 | 5.0 | 43510 | 0.0551 |
| 0.0246 | 6.0 | 52212 | 0.0588 |
| 0.0284 | 7.0 | 60914 | 0.0557 |
| 0.0303 | 8.0 | 69616 | 0.0543 |
| 0.0239 | 9.0 | 78318 | 0.0634 |
| 0.0207 | 10.0 | 87020 | 0.0617 |
Framework versions
- Transformers 4.21.1
- Pytorch 1.12.1+cu113
- Datasets 2.4.0
- Tokenizers 0.12.1