small-mlm-glue-cola-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: nan
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 |
|---|---|---|---|
| 6.3833 | 0.47 | 500 | 6.0687 |
| 5.7664 | 0.94 | 1000 | 5.7078 |
| 5.4733 | 1.4 | 1500 | 5.6928 |
| 5.2967 | 1.87 | 2000 | 5.6740 |
| 5.0316 | 2.34 | 2500 | 5.5465 |
| 5.0508 | 2.81 | 3000 | nan |
| 5.003 | 3.27 | 3500 | 5.4421 |
| 4.7143 | 3.74 | 4000 | 5.3365 |
| 4.7795 | 4.21 | 4500 | 5.2115 |
| 4.6303 | 4.68 | 5000 | 5.0677 |
| 4.5267 | 5.14 | 5500 | 5.2197 |
| 4.5582 | 5.61 | 6000 | 5.0485 |
| 4.528 | 6.08 | 6500 | 5.1092 |
| 4.2633 | 6.55 | 7000 | 5.1459 |
| 4.2774 | 7.02 | 7500 | nan |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2