small-mlm-glue-qqp-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: 5.0065
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 |
|---|---|---|---|
| 6.3631 | 0.4 | 500 | 5.9145 |
| 5.6422 | 0.8 | 1000 | 5.8224 |
| 5.4368 | 1.2 | 1500 | 5.6172 |
| 5.1539 | 1.6 | 2000 | 5.4872 |
| 5.0641 | 2.0 | 2500 | 5.5369 |
| 4.9495 | 2.4 | 3000 | 5.3466 |
| 4.8947 | 2.8 | 3500 | 5.4592 |
| 4.9081 | 3.2 | 4000 | 5.3328 |
| 4.7214 | 3.6 | 4500 | 5.3746 |
| 4.7341 | 4.0 | 5000 | 5.3417 |
| 4.6482 | 4.4 | 5500 | 5.2731 |
| 4.628 | 4.8 | 6000 | 5.2716 |
| 4.5801 | 5.2 | 6500 | 5.1364 |
| 4.4967 | 5.6 | 7000 | 5.2167 |
| 4.4984 | 6.0 | 7500 | 5.2133 |
| 4.4255 | 6.4 | 8000 | 5.1228 |
| 4.4459 | 6.8 | 8500 | 5.1664 |
| 4.3732 | 7.2 | 9000 | 5.0800 |
| 4.2546 | 7.6 | 9500 | 5.0616 |
| 4.351 | 8.0 | 10000 | 5.1500 |
| 4.2365 | 8.4 | 10500 | 5.0903 |
| 4.2224 | 8.8 | 11000 | 5.0041 |
| 4.2549 | 9.2 | 11500 | 5.0711 |
| 4.1108 | 9.6 | 12000 | 5.1525 |
| 4.1366 | 10.0 | 12500 | 5.0065 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2