<!-- 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_L10_H128_A2_massive
This model is a fine-tuned version of gokuls/HBERTv1_48_L10_H128_A2 on the massive dataset. It achieves the following results on the evaluation set:
- Loss: 1.0992
- Accuracy: 0.7550
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 |
---|---|---|---|---|
3.7898 | 1.0 | 180 | 3.3951 | 0.2120 |
3.0906 | 2.0 | 360 | 2.7701 | 0.3261 |
2.5803 | 3.0 | 540 | 2.3090 | 0.4235 |
2.1582 | 4.0 | 720 | 1.9427 | 0.5027 |
1.824 | 5.0 | 900 | 1.6937 | 0.5898 |
1.5669 | 6.0 | 1080 | 1.5220 | 0.6296 |
1.3731 | 7.0 | 1260 | 1.3841 | 0.6690 |
1.2237 | 8.0 | 1440 | 1.2811 | 0.6965 |
1.1037 | 9.0 | 1620 | 1.2167 | 0.7127 |
1.0132 | 10.0 | 1800 | 1.1843 | 0.7177 |
0.9387 | 11.0 | 1980 | 1.1377 | 0.7275 |
0.8742 | 12.0 | 2160 | 1.1167 | 0.7437 |
0.8335 | 13.0 | 2340 | 1.1076 | 0.7486 |
0.7923 | 14.0 | 2520 | 1.0992 | 0.7550 |
0.7673 | 15.0 | 2700 | 1.0983 | 0.7516 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.0