<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
distilbert_add_GLUE_Experiment_mnli_256
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: 0.9622
- Accuracy: 0.5456
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 |
|---|---|---|---|---|
| 1.0905 | 1.0 | 1534 | 1.0415 | 0.4633 |
| 1.0344 | 2.0 | 3068 | 1.0571 | 0.4569 |
| 1.0137 | 3.0 | 4602 | 1.0080 | 0.4888 |
| 0.9902 | 4.0 | 6136 | 0.9876 | 0.5090 |
| 0.9733 | 5.0 | 7670 | 0.9744 | 0.5134 |
| 0.9596 | 6.0 | 9204 | 0.9726 | 0.5138 |
| 0.9484 | 7.0 | 10738 | 0.9684 | 0.5233 |
| 0.938 | 8.0 | 12272 | 0.9715 | 0.5150 |
| 0.9279 | 9.0 | 13806 | 0.9707 | 0.5237 |
| 0.9188 | 10.0 | 15340 | 0.9580 | 0.5319 |
| 0.9102 | 11.0 | 16874 | 0.9561 | 0.5396 |
| 0.9011 | 12.0 | 18408 | 0.9594 | 0.5368 |
| 0.893 | 13.0 | 19942 | 0.9641 | 0.5345 |
| 0.8845 | 14.0 | 21476 | 0.9671 | 0.5367 |
| 0.8751 | 15.0 | 23010 | 0.9553 | 0.5388 |
| 0.8662 | 16.0 | 24544 | 0.9601 | 0.5447 |
| 0.8573 | 17.0 | 26078 | 0.9519 | 0.5473 |
| 0.849 | 18.0 | 27612 | 0.9814 | 0.5410 |
| 0.8404 | 19.0 | 29146 | 0.9733 | 0.5464 |
| 0.832 | 20.0 | 30680 | 0.9671 | 0.5492 |
| 0.8233 | 21.0 | 32214 | 0.9864 | 0.5509 |
| 0.8144 | 22.0 | 33748 | 0.9925 | 0.5536 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.8.0
- Tokenizers 0.13.2
