distilbert_add_GLUE_Experiment_logit_kd_mnli_192
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: 0.5625
- Accuracy: 0.5101
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 |
|---|---|---|---|---|
| 0.6241 | 1.0 | 1534 | 0.6196 | 0.3488 |
| 0.6176 | 2.0 | 3068 | 0.6093 | 0.3740 |
| 0.6054 | 3.0 | 4602 | 0.5983 | 0.4118 |
| 0.5943 | 4.0 | 6136 | 0.6006 | 0.4191 |
| 0.5849 | 5.0 | 7670 | 0.5842 | 0.4655 |
| 0.5752 | 6.0 | 9204 | 0.5992 | 0.4738 |
| 0.5671 | 7.0 | 10738 | 0.5735 | 0.4934 |
| 0.5615 | 8.0 | 12272 | 0.5844 | 0.4969 |
| 0.5577 | 9.0 | 13806 | 0.5643 | 0.5103 |
| 0.5539 | 10.0 | 15340 | 0.5796 | 0.4902 |
| 0.5509 | 11.0 | 16874 | 0.5696 | 0.5103 |
| 0.5484 | 12.0 | 18408 | 0.5829 | 0.5041 |
| 0.5455 | 13.0 | 19942 | 0.5718 | 0.5143 |
| 0.5428 | 14.0 | 21476 | 0.5722 | 0.5129 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2