<!-- 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. -->
hBERTv2_new_pretrain_w_init_48_ver2_stsb
This model is a fine-tuned version of gokuls/bert_12_layer_model_v2_complete_training_new_wt_init_48 on the GLUE STSB dataset. It achieves the following results on the evaluation set:
- Loss: 2.2194
- Pearson: 0.2187
- Spearmanr: 0.1976
- Combined Score: 0.2081
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: 4e-05
- train_batch_size: 64
- eval_batch_size: 64
- seed: 10
- 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 | Pearson | Spearmanr | Combined Score |
|---|---|---|---|---|---|---|
| 2.3584 | 1.0 | 90 | 2.3085 | 0.1702 | 0.1471 | 0.1586 |
| 2.0513 | 2.0 | 180 | 2.4060 | 0.1479 | 0.1342 | 0.1411 |
| 1.9851 | 3.0 | 270 | 2.4888 | 0.0897 | 0.1163 | 0.1030 |
| 1.8287 | 4.0 | 360 | 2.7571 | 0.1643 | 0.1827 | 0.1735 |
| 1.6845 | 5.0 | 450 | 2.2194 | 0.2187 | 0.1976 | 0.2081 |
| 1.6892 | 6.0 | 540 | 2.4431 | 0.1882 | 0.1858 | 0.1870 |
| 1.5272 | 7.0 | 630 | 2.6124 | 0.1433 | 0.1572 | 0.1503 |
| 1.402 | 8.0 | 720 | 2.8100 | 0.1605 | 0.1671 | 0.1638 |
| 1.3122 | 9.0 | 810 | 2.7081 | 0.1298 | 0.1428 | 0.1363 |
| 1.187 | 10.0 | 900 | 2.8638 | 0.1724 | 0.1825 | 0.1775 |
Framework versions
- Transformers 4.34.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.14.5
- Tokenizers 0.14.1
