tiny-vanilla-target-glue-qqp
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: 0.4162
- Accuracy: 0.7951
- F1: 0.7610
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 | Accuracy | F1 |
|---|---|---|---|---|---|
| 0.5864 | 0.04 | 500 | 0.5228 | 0.7257 | 0.6710 |
| 0.5173 | 0.09 | 1000 | 0.4944 | 0.7372 | 0.7000 |
| 0.5005 | 0.13 | 1500 | 0.4983 | 0.7317 | 0.7096 |
| 0.4853 | 0.18 | 2000 | 0.4763 | 0.7451 | 0.7176 |
| 0.474 | 0.22 | 2500 | 0.4644 | 0.7546 | 0.7240 |
| 0.4584 | 0.26 | 3000 | 0.4570 | 0.7617 | 0.7321 |
| 0.4584 | 0.31 | 3500 | 0.4513 | 0.7640 | 0.7348 |
| 0.4531 | 0.35 | 4000 | 0.4587 | 0.7587 | 0.7345 |
| 0.4536 | 0.4 | 4500 | 0.4523 | 0.7627 | 0.7383 |
| 0.4444 | 0.44 | 5000 | 0.4282 | 0.7847 | 0.7467 |
| 0.4323 | 0.48 | 5500 | 0.4415 | 0.7718 | 0.7445 |
| 0.4315 | 0.53 | 6000 | 0.4130 | 0.7969 | 0.7556 |
| 0.4275 | 0.57 | 6500 | 0.4339 | 0.7791 | 0.7506 |
| 0.4272 | 0.62 | 7000 | 0.4127 | 0.7949 | 0.7568 |
| 0.4182 | 0.66 | 7500 | 0.4162 | 0.7951 | 0.7610 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2