roberta-base-cola
This model is a fine-tuned version of roberta-base on the GLUE COLA dataset. It achieves the following results on the evaluation set:
- Loss: 1.0571
- Matthews Correlation: 0.6232
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.06
- num_epochs: 10.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
|---|---|---|---|---|
| 0.5497 | 1.0 | 535 | 0.5504 | 0.4613 |
| 0.3786 | 2.0 | 1070 | 0.4850 | 0.5470 |
| 0.2733 | 3.0 | 1605 | 0.5036 | 0.5792 |
| 0.2204 | 4.0 | 2140 | 0.5532 | 0.6139 |
| 0.164 | 5.0 | 2675 | 0.9516 | 0.5934 |
| 0.1351 | 6.0 | 3210 | 0.9051 | 0.5754 |
| 0.1065 | 7.0 | 3745 | 0.9006 | 0.6161 |
| 0.0874 | 8.0 | 4280 | 0.9457 | 0.6157 |
| 0.0579 | 9.0 | 4815 | 1.0372 | 0.6007 |
| 0.0451 | 10.0 | 5350 | 1.0571 | 0.6232 |
Framework versions
- Transformers 4.20.0.dev0
- Pytorch 1.11.0+cu113
- Datasets 2.1.0
- Tokenizers 0.12.1