Bert_tinybert-distilled
This model is a fine-tuned version of google/bert_uncased_L-2_H-128_A-2 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4565
- Accuracy: 0.7973
- Auc: 0.8745
- Mcc: 0.6019
- Pr Auc: 0.8517
- Bac: 0.7973
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: 0.0003098011180790181
- train_batch_size: 32
- eval_batch_size: 32
- seed: 33
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 9
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Auc | Mcc | Pr Auc | Bac |
|---|---|---|---|---|---|---|---|---|
| 0.6039 | 1.0 | 172 | 0.5328 | 0.7033 | 0.8094 | 0.4152 | 0.7669 | 0.7033 |
| 0.5426 | 2.0 | 344 | 0.4892 | 0.78 | 0.8386 | 0.5777 | 0.7876 | 0.78 |
| 0.5225 | 3.0 | 516 | 0.4794 | 0.7847 | 0.8532 | 0.5774 | 0.8195 | 0.7847 |
| 0.5069 | 4.0 | 688 | 0.4593 | 0.7973 | 0.8656 | 0.5988 | 0.8310 | 0.7973 |
| 0.4971 | 5.0 | 860 | 0.4636 | 0.8 | 0.8696 | 0.6114 | 0.8421 | 0.8 |
| 0.479 | 6.0 | 1032 | 0.4567 | 0.802 | 0.8744 | 0.6146 | 0.8520 | 0.802 |
| 0.4733 | 7.0 | 1204 | 0.4559 | 0.8007 | 0.8706 | 0.6054 | 0.8510 | 0.8007 |
| 0.4647 | 8.0 | 1376 | 0.4525 | 0.8033 | 0.8733 | 0.6108 | 0.8499 | 0.8033 |
| 0.4646 | 9.0 | 1548 | 0.4565 | 0.7973 | 0.8745 | 0.6019 | 0.8517 | 0.7973 |
Framework versions
- Transformers 4.43.2
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1