<!-- 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_L12_H768_A12_massive
This model is a fine-tuned version of gokuls/HBERTv1_48_L12_H768_A12 on the massive dataset. It achieves the following results on the evaluation set:
- Loss: 0.8779
- Accuracy: 0.8598
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 |
|---|---|---|---|---|
| 1.9138 | 1.0 | 180 | 0.9291 | 0.7496 |
| 0.8364 | 2.0 | 360 | 0.7064 | 0.8214 |
| 0.5786 | 3.0 | 540 | 0.7115 | 0.8214 |
| 0.4285 | 4.0 | 720 | 0.6551 | 0.8357 |
| 0.3299 | 5.0 | 900 | 0.6935 | 0.8337 |
| 0.2483 | 6.0 | 1080 | 0.6726 | 0.8372 |
| 0.1857 | 7.0 | 1260 | 0.6919 | 0.8480 |
| 0.1393 | 8.0 | 1440 | 0.7360 | 0.8460 |
| 0.1117 | 9.0 | 1620 | 0.7913 | 0.8470 |
| 0.0748 | 10.0 | 1800 | 0.7926 | 0.8495 |
| 0.0536 | 11.0 | 1980 | 0.8390 | 0.8446 |
| 0.0351 | 12.0 | 2160 | 0.8506 | 0.8549 |
| 0.0233 | 13.0 | 2340 | 0.8834 | 0.8529 |
| 0.0121 | 14.0 | 2520 | 0.8779 | 0.8598 |
| 0.0065 | 15.0 | 2700 | 0.8833 | 0.8588 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1
