bert_uncased_L-12_H-128_A-2_massive
This model is a fine-tuned version of google/bert_uncased_L-12_H-128_A-2 on the massive dataset. It achieves the following results on the evaluation set:
- Loss: 1.4592
- Accuracy: 0.7452
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 |
|---|---|---|---|---|
| 3.7918 | 1.0 | 180 | 3.4894 | 0.3025 |
| 3.2847 | 2.0 | 360 | 3.0212 | 0.3748 |
| 2.884 | 3.0 | 540 | 2.6641 | 0.4889 |
| 2.568 | 4.0 | 720 | 2.3904 | 0.5888 |
| 2.3179 | 5.0 | 900 | 2.1799 | 0.6439 |
| 2.1182 | 6.0 | 1080 | 2.0118 | 0.6714 |
| 1.9499 | 7.0 | 1260 | 1.8764 | 0.7004 |
| 1.816 | 8.0 | 1440 | 1.7611 | 0.7103 |
| 1.7109 | 9.0 | 1620 | 1.6734 | 0.7231 |
| 1.6164 | 10.0 | 1800 | 1.6047 | 0.7309 |
| 1.5424 | 11.0 | 1980 | 1.5481 | 0.7324 |
| 1.4871 | 12.0 | 2160 | 1.5112 | 0.7393 |
| 1.4401 | 13.0 | 2340 | 1.4808 | 0.7437 |
| 1.4179 | 14.0 | 2520 | 1.4667 | 0.7432 |
| 1.4025 | 15.0 | 2700 | 1.4592 | 0.7452 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1