tiny-mlm-glue-mnli-custom-tokenizer
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: 6.1721
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 |
|---|---|---|---|
| 7.8162 | 0.4 | 500 | 7.1032 |
| 6.9567 | 0.8 | 1000 | 7.0697 |
| 6.8563 | 1.2 | 1500 | 7.0460 |
| 6.7685 | 1.6 | 2000 | 7.0131 |
| 6.6897 | 2.0 | 2500 | 6.9769 |
| 6.5455 | 2.4 | 3000 | 6.9249 |
| 6.482 | 2.8 | 3500 | 6.8552 |
| 6.4153 | 3.2 | 4000 | 6.8445 |
| 6.38 | 3.6 | 4500 | 6.7803 |
| 6.4066 | 4.0 | 5000 | 6.8070 |
| 6.2854 | 4.4 | 5500 | 6.7329 |
| 6.2966 | 4.8 | 6000 | 6.7094 |
| 6.1244 | 5.2 | 6500 | 6.6476 |
| 6.1276 | 5.6 | 7000 | 6.6118 |
| 6.0685 | 6.0 | 7500 | 6.5714 |
| 5.98 | 6.4 | 8000 | 6.5522 |
| 6.0174 | 6.8 | 8500 | 6.5093 |
| 5.9451 | 7.2 | 9000 | 6.4866 |
| 5.9681 | 7.6 | 9500 | 6.5238 |
| 5.9246 | 8.0 | 10000 | 6.5340 |
| 5.9219 | 8.4 | 10500 | 6.4727 |
| 5.8812 | 8.8 | 11000 | 6.4483 |
| 5.7815 | 9.2 | 11500 | 6.4402 |
| 5.7938 | 9.6 | 12000 | 6.4124 |
| 5.7934 | 10.0 | 12500 | 6.3908 |
| 5.7332 | 10.4 | 13000 | 6.3861 |
| 5.7628 | 10.8 | 13500 | 6.3638 |
| 5.7259 | 11.2 | 14000 | 6.3345 |
| 5.7505 | 11.6 | 14500 | 6.3117 |
| 5.6441 | 12.0 | 15000 | 6.3118 |
| 5.7058 | 12.4 | 15500 | 6.3116 |
| 5.6017 | 12.8 | 16000 | 6.2728 |
| 5.6424 | 13.2 | 16500 | 6.2790 |
| 5.5799 | 13.6 | 17000 | 6.3034 |
| 5.5625 | 14.0 | 17500 | 6.2580 |
| 5.6015 | 14.4 | 18000 | 6.2607 |
| 5.4884 | 14.8 | 18500 | 6.2535 |
| 5.5117 | 15.2 | 19000 | 6.1960 |
| 5.4919 | 15.6 | 19500 | 6.1907 |
| 5.4624 | 16.0 | 20000 | 6.1838 |
| 5.4721 | 16.4 | 20500 | 6.1461 |
| 5.4833 | 16.8 | 21000 | 6.1251 |
| 5.4404 | 17.2 | 21500 | 6.1725 |
| 5.4487 | 17.6 | 22000 | 6.1417 |
| 5.4499 | 18.0 | 22500 | 6.1721 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2