distilbert_sa_GLUE_Experiment_logit_kd_mrpc_256
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: 0.5199
- Accuracy: 0.3284
- F1: 0.0616
- Combined Score: 0.1950
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 |
|---|---|---|---|---|---|---|
| 0.5375 | 1.0 | 15 | 0.5292 | 0.3162 | 0.0 | 0.1581 |
| 0.5305 | 2.0 | 30 | 0.5292 | 0.3162 | 0.0 | 0.1581 |
| 0.5294 | 3.0 | 45 | 0.5293 | 0.3162 | 0.0 | 0.1581 |
| 0.5283 | 4.0 | 60 | 0.5284 | 0.3162 | 0.0 | 0.1581 |
| 0.5258 | 5.0 | 75 | 0.5260 | 0.3162 | 0.0 | 0.1581 |
| 0.519 | 6.0 | 90 | 0.5199 | 0.3284 | 0.0616 | 0.1950 |
| 0.5036 | 7.0 | 105 | 0.5200 | 0.3848 | 0.2462 | 0.3155 |
| 0.4916 | 8.0 | 120 | 0.5226 | 0.4167 | 0.3239 | 0.3703 |
| 0.4725 | 9.0 | 135 | 0.5298 | 0.4289 | 0.3581 | 0.3935 |
| 0.4537 | 10.0 | 150 | 0.5333 | 0.6152 | 0.6736 | 0.6444 |
| 0.4382 | 11.0 | 165 | 0.5450 | 0.6201 | 0.6906 | 0.6554 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2