distilbert_sa_GLUE_Experiment_data_aug_qqp_384
This model is a fine-tuned version of distilbert-base-uncased on the GLUE QQP dataset. It achieves the following results on the evaluation set:
- Loss: 0.5363
- Accuracy: 0.7995
- F1: 0.7338
- Combined Score: 0.7666
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: 5e-05
- train_batch_size: 256
- eval_batch_size: 256
- seed: 10
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Accuracy | Combined Score | F1 | Validation Loss |
|---|---|---|---|---|---|---|
| 0.3538 | 1.0 | 29671 | 0.7995 | 0.7666 | 0.7338 | 0.5363 |
| 0.1571 | 2.0 | 59342 | 0.7215 | 0.8000 | 0.7396 | 0.7698 |
| 0.0894 | 3.0 | 89013 | 0.7922 | 0.7998 | 0.7407 | 0.7702 |
| 0.0596 | 4.0 | 118684 | 0.8829 | 0.8045 | 0.7399 | 0.7722 |
| 0.0433 | 5.0 | 148355 | 0.8505 | 0.8110 | 0.7443 | 0.7777 |
| 0.0334 | 6.0 | 178026 | 1.0843 | 0.8047 | 0.7446 | 0.7746 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2