distilbert_sa_GLUE_Experiment_logit_kd_data_aug_qqp_256
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: 0.7043
- Accuracy: 0.6343
- F1: 0.0148
- Combined Score: 0.3245
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 | Validation Loss | Accuracy | F1 | Combined Score |
|---|---|---|---|---|---|---|
| 0.8369 | 1.0 | 29671 | 0.7043 | 0.6343 | 0.0148 | 0.3245 |
| 0.7448 | 2.0 | 59342 | 0.7161 | 0.6355 | 0.0216 | 0.3286 |
| 0.7106 | 3.0 | 89013 | 0.7067 | 0.6466 | 0.0843 | 0.3655 |
| 0.6924 | 4.0 | 118684 | 0.7200 | 0.6401 | 0.0477 | 0.3439 |
| 0.6812 | 5.0 | 148355 | 0.7109 | 0.6424 | 0.0609 | 0.3517 |
| 0.6734 | 6.0 | 178026 | 0.7092 | 0.6440 | 0.0696 | 0.3568 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2