distilbert_sa_GLUE_Experiment_logit_kd_mrpc_192
This model is a fine-tuned version of distilbert-base-uncased on the GLUE MRPC dataset. It achieves the following results on the evaluation set:
- Loss: 0.5189
- Accuracy: 0.3382
- F1: 0.0816
- Combined Score: 0.2099
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 |
|---|---|---|---|---|---|---|
| 0.5329 | 1.0 | 15 | 0.5292 | 0.3162 | 0.0 | 0.1581 |
| 0.5309 | 2.0 | 30 | 0.5294 | 0.3162 | 0.0 | 0.1581 |
| 0.5291 | 3.0 | 45 | 0.5292 | 0.3162 | 0.0 | 0.1581 |
| 0.5286 | 4.0 | 60 | 0.5288 | 0.3162 | 0.0 | 0.1581 |
| 0.5269 | 5.0 | 75 | 0.5277 | 0.3162 | 0.0 | 0.1581 |
| 0.5255 | 6.0 | 90 | 0.5246 | 0.3162 | 0.0 | 0.1581 |
| 0.5157 | 7.0 | 105 | 0.5189 | 0.3382 | 0.0816 | 0.2099 |
| 0.5037 | 8.0 | 120 | 0.5221 | 0.3284 | 0.0486 | 0.1885 |
| 0.4859 | 9.0 | 135 | 0.5277 | 0.4681 | 0.4151 | 0.4416 |
| 0.4683 | 10.0 | 150 | 0.5407 | 0.5882 | 0.6364 | 0.6123 |
| 0.4558 | 11.0 | 165 | 0.5487 | 0.4951 | 0.4772 | 0.4861 |
| 0.4439 | 12.0 | 180 | 0.5611 | 0.5319 | 0.5527 | 0.5423 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2