distilbert_sa_GLUE_Experiment_logit_kd_mnli_96
This model is a fine-tuned version of distilbert-base-uncased on the GLUE MNLI dataset. It achieves the following results on the evaluation set:
- Loss: 0.5438
- Accuracy: 0.5431
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 |
|---|---|---|---|---|
| 0.6023 | 1.0 | 1534 | 0.5718 | 0.4960 |
| 0.5673 | 2.0 | 3068 | 0.5547 | 0.5184 |
| 0.5555 | 3.0 | 4602 | 0.5505 | 0.5278 |
| 0.5481 | 4.0 | 6136 | 0.5466 | 0.5381 |
| 0.5426 | 5.0 | 7670 | 0.5454 | 0.5403 |
| 0.5382 | 6.0 | 9204 | 0.5454 | 0.5354 |
| 0.5341 | 7.0 | 10738 | 0.5452 | 0.5344 |
| 0.5308 | 8.0 | 12272 | 0.5428 | 0.5410 |
| 0.5271 | 9.0 | 13806 | 0.5460 | 0.5451 |
| 0.5239 | 10.0 | 15340 | 0.5450 | 0.5462 |
| 0.5209 | 11.0 | 16874 | 0.5447 | 0.5449 |
| 0.5179 | 12.0 | 18408 | 0.5452 | 0.5475 |
| 0.5152 | 13.0 | 19942 | 0.5495 | 0.5454 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2