<!-- 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_H512_A8_massive
This model is a fine-tuned version of gokuls/HBERTv1_48_L12_H512_A8 on the massive dataset. It achieves the following results on the evaluation set:
- Loss: 0.8114
- Accuracy: 0.8564
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 |
|---|---|---|---|---|
| 2.5697 | 1.0 | 180 | 1.3233 | 0.6562 |
| 1.0695 | 2.0 | 360 | 0.8666 | 0.7654 |
| 0.7172 | 3.0 | 540 | 0.7910 | 0.7964 |
| 0.5436 | 4.0 | 720 | 0.7059 | 0.8214 |
| 0.4136 | 5.0 | 900 | 0.7533 | 0.8141 |
| 0.3215 | 6.0 | 1080 | 0.6837 | 0.8396 |
| 0.2429 | 7.0 | 1260 | 0.7176 | 0.8269 |
| 0.195 | 8.0 | 1440 | 0.7018 | 0.8451 |
| 0.1459 | 9.0 | 1620 | 0.7386 | 0.8446 |
| 0.112 | 10.0 | 1800 | 0.7815 | 0.8451 |
| 0.0806 | 11.0 | 1980 | 0.8076 | 0.8406 |
| 0.0587 | 12.0 | 2160 | 0.7878 | 0.8559 |
| 0.0424 | 13.0 | 2340 | 0.8044 | 0.8510 |
| 0.0272 | 14.0 | 2520 | 0.8114 | 0.8564 |
| 0.0192 | 15.0 | 2700 | 0.8199 | 0.8549 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.0
