bert_uncased_L-12_H-256_A-4_massive
This model is a fine-tuned version of google/bert_uncased_L-12_H-256_A-4 on the massive dataset. It achieves the following results on the evaluation set:
- Loss: 0.6679
- Accuracy: 0.8618
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: 64
- eval_batch_size: 64
- seed: 33
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 15
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 3.3504 | 1.0 | 180 | 2.5753 | 0.5726 |
| 2.2201 | 2.0 | 360 | 1.7801 | 0.7280 |
| 1.5878 | 3.0 | 540 | 1.3596 | 0.7747 |
| 1.218 | 4.0 | 720 | 1.1112 | 0.8032 |
| 0.9707 | 5.0 | 900 | 0.9673 | 0.8111 |
| 0.7935 | 6.0 | 1080 | 0.8617 | 0.8313 |
| 0.6704 | 7.0 | 1260 | 0.8082 | 0.8455 |
| 0.572 | 8.0 | 1440 | 0.7746 | 0.8446 |
| 0.5015 | 9.0 | 1620 | 0.7387 | 0.8500 |
| 0.4434 | 10.0 | 1800 | 0.7024 | 0.8534 |
| 0.3947 | 11.0 | 1980 | 0.7013 | 0.8549 |
| 0.362 | 12.0 | 2160 | 0.6884 | 0.8544 |
| 0.3365 | 13.0 | 2340 | 0.6821 | 0.8549 |
| 0.3172 | 14.0 | 2520 | 0.6704 | 0.8593 |
| 0.3067 | 15.0 | 2700 | 0.6679 | 0.8618 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1