sa_BERT_24_qnli
This model is a fine-tuned version of gokuls/bert_base_24 on the GLUE QNLI dataset. It achieves the following results on the evaluation set:
- Loss: 0.6492
- Accuracy: 0.6218
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: 4e-05
- train_batch_size: 96
- eval_batch_size: 96
- 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
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.6745 | 1.0 | 1092 | 0.6686 | 0.5817 |
| 0.6338 | 2.0 | 2184 | 0.6492 | 0.6218 |
| 0.5909 | 3.0 | 3276 | 0.6560 | 0.6251 |
| 0.5407 | 4.0 | 4368 | 0.7246 | 0.6269 |
| 0.4732 | 5.0 | 5460 | 0.6612 | 0.6421 |
| 0.3999 | 6.0 | 6552 | 0.7506 | 0.6410 |
| 0.3203 | 7.0 | 7644 | 0.9162 | 0.6306 |
Framework versions
- Transformers 4.30.2
- Pytorch 1.14.0a0+410ce96
- Datasets 2.13.0
- Tokenizers 0.13.3