<!-- 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. -->
bert-base-uncased-stsb
This model is a fine-tuned version of bert-base-uncased on the GLUE STSB dataset. It achieves the following results on the evaluation set:
- Loss: 0.4676
- Pearson: 0.8901
- Spearmanr: 0.8872
- Combined Score: 0.8887
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: 128
- eval_batch_size: 128
- seed: 10
- distributed_type: multi-GPU
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 50
Training results
Training Loss | Epoch | Step | Validation Loss | Pearson | Spearmanr | Combined Score |
---|---|---|---|---|---|---|
2.3939 | 1.0 | 45 | 0.7358 | 0.8686 | 0.8653 | 0.8669 |
0.5084 | 2.0 | 90 | 0.4959 | 0.8835 | 0.8799 | 0.8817 |
0.3332 | 3.0 | 135 | 0.5002 | 0.8846 | 0.8815 | 0.8830 |
0.2202 | 4.0 | 180 | 0.4962 | 0.8854 | 0.8827 | 0.8840 |
0.1642 | 5.0 | 225 | 0.4848 | 0.8864 | 0.8839 | 0.8852 |
0.1312 | 6.0 | 270 | 0.4987 | 0.8872 | 0.8866 | 0.8869 |
0.1057 | 7.0 | 315 | 0.4840 | 0.8895 | 0.8848 | 0.8871 |
0.0935 | 8.0 | 360 | 0.4753 | 0.8887 | 0.8840 | 0.8863 |
0.0835 | 9.0 | 405 | 0.4676 | 0.8901 | 0.8872 | 0.8887 |
0.0749 | 10.0 | 450 | 0.4808 | 0.8901 | 0.8867 | 0.8884 |
0.0625 | 11.0 | 495 | 0.4760 | 0.8893 | 0.8857 | 0.8875 |
0.0607 | 12.0 | 540 | 0.5113 | 0.8899 | 0.8859 | 0.8879 |
0.0564 | 13.0 | 585 | 0.4918 | 0.8900 | 0.8860 | 0.8880 |
0.0495 | 14.0 | 630 | 0.4749 | 0.8905 | 0.8868 | 0.8887 |
0.0446 | 15.0 | 675 | 0.4889 | 0.8888 | 0.8856 | 0.8872 |
0.045 | 16.0 | 720 | 0.4680 | 0.8918 | 0.8889 | 0.8904 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2