tiny-vanilla-target-conll2003
This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1431
- Precision: 0.7507
- Recall: 0.8177
- F1: 0.7828
- Accuracy: 0.9581
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
- training_steps: 5000
Training results
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|---|---|---|---|---|---|---|---|
| 0.7673 | 1.14 | 500 | 0.4291 | 0.4793 | 0.5160 | 0.4970 | 0.8920 |
| 0.3746 | 2.28 | 1000 | 0.2869 | 0.5976 | 0.6572 | 0.6260 | 0.9256 |
| 0.2869 | 3.42 | 1500 | 0.2292 | 0.6411 | 0.7184 | 0.6776 | 0.9370 |
| 0.236 | 4.56 | 2000 | 0.1988 | 0.6805 | 0.7516 | 0.7143 | 0.9438 |
| 0.2026 | 5.69 | 2500 | 0.1772 | 0.7047 | 0.7718 | 0.7367 | 0.9482 |
| 0.1798 | 6.83 | 3000 | 0.1649 | 0.7179 | 0.7864 | 0.7506 | 0.9514 |
| 0.158 | 7.97 | 3500 | 0.1559 | 0.7256 | 0.7987 | 0.7604 | 0.9543 |
| 0.1415 | 9.11 | 4000 | 0.1500 | 0.7379 | 0.8034 | 0.7693 | 0.9563 |
| 0.127 | 10.25 | 4500 | 0.1462 | 0.7532 | 0.8134 | 0.7821 | 0.9573 |
| 0.1173 | 11.39 | 5000 | 0.1431 | 0.7507 | 0.8177 | 0.7828 | 0.9581 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2