bert-base-uncased-finetuned-cola-learning_rate-2e-05
This model is a fine-tuned version of bert-base-uncased on the glue dataset. It achieves the following results on the evaluation set:
- Loss: 0.4480
- Matthews Correlation: 0.5892
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
Training results
Training Loss | Epoch | Step | Validation Loss | Matthews Correlation |
---|---|---|---|---|
0.5052 | 1.0 | 535 | 0.5532 | 0.5030 |
0.3006 | 2.0 | 1070 | 0.4480 | 0.5892 |
0.1918 | 3.0 | 1605 | 0.7164 | 0.5340 |
0.138 | 4.0 | 2140 | 0.8575 | 0.5570 |
0.0866 | 5.0 | 2675 | 1.1483 | 0.5211 |
0.0652 | 6.0 | 3210 | 0.9938 | 0.5816 |
0.046 | 7.0 | 3745 | 1.1453 | 0.5739 |
0.0314 | 8.0 | 4280 | 1.3524 | 0.5573 |
0.0212 | 9.0 | 4815 | 1.4664 | 0.5573 |
0.0203 | 10.0 | 5350 | 1.4505 | 0.5679 |
Framework versions
- Transformers 4.28.1
- Pytorch 2.0.0+cu118
- Datasets 2.12.0
- Tokenizers 0.13.3
- Downloads last month
- 1
Dataset used to train cansurav/bert-base-uncased-finetuned-cola-learning_rate-2e-05
Evaluation results
- Matthews Correlation on gluevalidation set self-reported0.589