distilbert_add_GLUE_Experiment_qqp_256
This model is a fine-tuned version of distilbert-base-uncased on the GLUE QQP dataset. It achieves the following results on the evaluation set:
- Loss: 0.4273
- Accuracy: 0.8136
- F1: 0.7425
- Combined Score: 0.7781
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: 256
- eval_batch_size: 256
- 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
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Combined Score |
|---|---|---|---|---|---|---|
| 0.5648 | 1.0 | 1422 | 0.5394 | 0.7316 | 0.6540 | 0.6928 |
| 0.5038 | 2.0 | 2844 | 0.5093 | 0.7496 | 0.6564 | 0.7030 |
| 0.4837 | 3.0 | 4266 | 0.4952 | 0.7623 | 0.6625 | 0.7124 |
| 0.4624 | 4.0 | 5688 | 0.4777 | 0.7739 | 0.6844 | 0.7292 |
| 0.4197 | 5.0 | 7110 | 0.4541 | 0.7925 | 0.6939 | 0.7432 |
| 0.3693 | 6.0 | 8532 | 0.4539 | 0.8027 | 0.7012 | 0.7519 |
| 0.3214 | 7.0 | 9954 | 0.4273 | 0.8136 | 0.7425 | 0.7781 |
| 0.2804 | 8.0 | 11376 | 0.4547 | 0.8187 | 0.7344 | 0.7765 |
| 0.2463 | 9.0 | 12798 | 0.4779 | 0.8227 | 0.7478 | 0.7852 |
| 0.2177 | 10.0 | 14220 | 0.5060 | 0.8256 | 0.7510 | 0.7883 |
| 0.1933 | 11.0 | 15642 | 0.5020 | 0.8272 | 0.7587 | 0.7929 |
| 0.1741 | 12.0 | 17064 | 0.5385 | 0.8304 | 0.7604 | 0.7954 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.8.0
- Tokenizers 0.13.2