distilbert_sa_GLUE_Experiment_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: 0.9073
- Accuracy: 0.5665
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 |
|---|---|---|---|---|
| 1.0297 | 1.0 | 1534 | 0.9819 | 0.5076 |
| 0.9555 | 2.0 | 3068 | 0.9376 | 0.5479 |
| 0.9202 | 3.0 | 4602 | 0.9282 | 0.5506 |
| 0.8951 | 4.0 | 6136 | 0.9175 | 0.5629 |
| 0.8759 | 5.0 | 7670 | 0.9040 | 0.5750 |
| 0.8587 | 6.0 | 9204 | 0.9110 | 0.5670 |
| 0.8422 | 7.0 | 10738 | 0.9196 | 0.5693 |
| 0.8261 | 8.0 | 12272 | 0.9521 | 0.5577 |
| 0.8114 | 9.0 | 13806 | 0.9293 | 0.5744 |
| 0.7967 | 10.0 | 15340 | 0.9075 | 0.5839 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.8.0
- Tokenizers 0.13.2