tiny-mlm-glue-wnli-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: 2.5277
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 |
|---|---|---|---|
| 6.6853 | 6.25 | 500 | 6.1267 |
| 5.7417 | 12.5 | 1000 | 5.9357 |
| 5.6014 | 18.75 | 1500 | 5.7487 |
| 5.4076 | 25.0 | 2000 | 5.2954 |
| 5.1906 | 31.25 | 2500 | 5.1483 |
| 4.9775 | 37.5 | 3000 | 5.2707 |
| 4.7267 | 43.75 | 3500 | 4.5667 |
| 4.4882 | 50.0 | 4000 | 4.7044 |
| 4.2548 | 56.25 | 4500 | 4.0688 |
| 4.0119 | 62.5 | 5000 | 3.8329 |
| 3.7807 | 68.75 | 5500 | 3.7323 |
| 3.5306 | 75.0 | 6000 | 3.2662 |
| 3.3206 | 81.25 | 6500 | 3.2223 |
| 3.1002 | 87.5 | 7000 | 3.1473 |
| 2.9093 | 93.75 | 7500 | 2.7988 |
| 2.7146 | 100.0 | 8000 | 2.3988 |
| 2.5981 | 106.25 | 8500 | 2.5495 |
| 2.4007 | 112.5 | 9000 | 2.4030 |
| 2.2446 | 118.75 | 9500 | 2.5277 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2