distilbert_sa_GLUE_Experiment_logit_kd_cola_96
This model is a fine-tuned version of distilbert-base-uncased on the GLUE COLA dataset. It achieves the following results on the evaluation set:
- Loss: 0.6770
- Matthews Correlation: 0.0438
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 | Matthews Correlation |
|---|---|---|---|---|
| 0.915 | 1.0 | 34 | 0.7697 | 0.0 |
| 0.8596 | 2.0 | 68 | 0.7301 | 0.0 |
| 0.826 | 3.0 | 102 | 0.7022 | 0.0 |
| 0.8072 | 4.0 | 136 | 0.6883 | 0.0 |
| 0.7996 | 5.0 | 170 | 0.6846 | 0.0 |
| 0.7958 | 6.0 | 204 | 0.6840 | 0.0 |
| 0.7977 | 7.0 | 238 | 0.6840 | 0.0 |
| 0.7973 | 8.0 | 272 | 0.6840 | 0.0 |
| 0.7954 | 9.0 | 306 | 0.6839 | 0.0 |
| 0.7963 | 10.0 | 340 | 0.6837 | 0.0 |
| 0.795 | 11.0 | 374 | 0.6817 | 0.0 |
| 0.7664 | 12.0 | 408 | 0.6770 | 0.0438 |
| 0.7144 | 13.0 | 442 | 0.6875 | 0.1060 |
| 0.6788 | 14.0 | 476 | 0.6928 | 0.0970 |
| 0.648 | 15.0 | 510 | 0.7124 | 0.1017 |
| 0.6288 | 16.0 | 544 | 0.7151 | 0.1005 |
| 0.613 | 17.0 | 578 | 0.7161 | 0.0812 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2