<!-- 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. -->
HBERTv1_48_L8_H768_A12_massive
This model is a fine-tuned version of gokuls/HBERTv1_48_L8_H768_A12 on the massive dataset. It achieves the following results on the evaluation set:
- Loss: 0.8748
- Accuracy: 0.8667
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 |
|---|---|---|---|---|
| 1.8167 | 1.0 | 180 | 0.8067 | 0.7846 |
| 0.7057 | 2.0 | 360 | 0.6923 | 0.8126 |
| 0.4613 | 3.0 | 540 | 0.6398 | 0.8406 |
| 0.3268 | 4.0 | 720 | 0.6181 | 0.8460 |
| 0.2406 | 5.0 | 900 | 0.6902 | 0.8382 |
| 0.1718 | 6.0 | 1080 | 0.6729 | 0.8539 |
| 0.1263 | 7.0 | 1260 | 0.7197 | 0.8470 |
| 0.0875 | 8.0 | 1440 | 0.7307 | 0.8623 |
| 0.06 | 9.0 | 1620 | 0.8116 | 0.8534 |
| 0.0494 | 10.0 | 1800 | 0.7662 | 0.8618 |
| 0.0271 | 11.0 | 1980 | 0.8777 | 0.8549 |
| 0.0189 | 12.0 | 2160 | 0.8954 | 0.8578 |
| 0.0113 | 13.0 | 2340 | 0.8748 | 0.8667 |
| 0.0055 | 14.0 | 2520 | 0.8829 | 0.8598 |
| 0.003 | 15.0 | 2700 | 0.8823 | 0.8613 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.0
