small-mlm-glue-mnli-custom-tokenizer
This model is a fine-tuned version of google/bert_uncased_L-4_H-512_A-8 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 5.6551
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
- num_epochs: 200
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 7.0308 | 0.4 | 500 | 6.6001 |
| 6.346 | 0.8 | 1000 | 6.3998 |
| 6.1061 | 1.2 | 1500 | 6.3170 |
| 5.9586 | 1.6 | 2000 | 6.2799 |
| 5.8773 | 2.0 | 2500 | 6.2034 |
| 5.7403 | 2.4 | 3000 | 6.1609 |
| 5.6602 | 2.8 | 3500 | 6.1113 |
| 5.5809 | 3.2 | 4000 | 6.1267 |
| 5.5663 | 3.6 | 4500 | 6.0647 |
| 5.6266 | 4.0 | 5000 | 6.1090 |
| 5.4756 | 4.4 | 5500 | 6.0302 |
| 5.4905 | 4.8 | 6000 | 6.0292 |
| 5.3179 | 5.2 | 6500 | 5.9758 |
| 5.3375 | 5.6 | 7000 | 6.0125 |
| 5.3035 | 6.0 | 7500 | 5.9495 |
| 5.1918 | 6.4 | 8000 | 5.9537 |
| 5.2499 | 6.8 | 8500 | 5.9100 |
| 5.1905 | 7.2 | 9000 | 5.8620 |
| 5.1787 | 7.6 | 9500 | 5.9296 |
| 5.1534 | 8.0 | 10000 | 5.9442 |
| 5.1396 | 8.4 | 10500 | 5.8609 |
| 5.1272 | 8.8 | 11000 | 5.8358 |
| 4.9615 | 9.2 | 11500 | 5.8617 |
| 5.0062 | 9.6 | 12000 | 5.8043 |
| 5.0131 | 10.0 | 12500 | 5.8119 |
| 4.9326 | 10.4 | 13000 | 5.7851 |
| 4.9655 | 10.8 | 13500 | 5.7792 |
| 4.9256 | 11.2 | 14000 | 5.7843 |
| 4.9195 | 11.6 | 14500 | 5.7652 |
| 4.8299 | 12.0 | 15000 | 5.7606 |
| 4.8748 | 12.4 | 15500 | 5.7577 |
| 4.7588 | 12.8 | 16000 | 5.7048 |
| 4.8185 | 13.2 | 16500 | 5.7245 |
| 4.7679 | 13.6 | 17000 | 5.7402 |
| 4.7377 | 14.0 | 17500 | 5.7034 |
| 4.7403 | 14.4 | 18000 | 5.7054 |
| 4.6628 | 14.8 | 18500 | 5.7203 |
| 4.6801 | 15.2 | 19000 | 5.6798 |
| 4.6014 | 15.6 | 19500 | 5.6931 |
| 4.618 | 16.0 | 20000 | 5.6620 |
| 4.6037 | 16.4 | 20500 | 5.6441 |
| 4.6004 | 16.8 | 21000 | 5.6262 |
| 4.5432 | 17.2 | 21500 | 5.6726 |
| 4.576 | 17.6 | 22000 | 5.6322 |
| 4.5568 | 18.0 | 22500 | 5.6551 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2