distilbert_sa_GLUE_Experiment_qqp_96
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: 0.4818
- Accuracy: 0.7757
- F1: 0.6672
- Combined Score: 0.7214
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 |
|---|---|---|---|---|---|---|
| 0.5528 | 1.0 | 1422 | 0.5115 | 0.7528 | 0.6284 | 0.6906 |
| 0.4965 | 2.0 | 2844 | 0.4960 | 0.7614 | 0.6420 | 0.7017 |
| 0.4769 | 3.0 | 4266 | 0.4904 | 0.7650 | 0.6382 | 0.7016 |
| 0.4619 | 4.0 | 5688 | 0.4901 | 0.7680 | 0.6526 | 0.7103 |
| 0.4489 | 5.0 | 7110 | 0.4844 | 0.7709 | 0.6663 | 0.7186 |
| 0.4366 | 6.0 | 8532 | 0.4860 | 0.7721 | 0.6712 | 0.7217 |
| 0.425 | 7.0 | 9954 | 0.4860 | 0.7747 | 0.6636 | 0.7192 |
| 0.414 | 8.0 | 11376 | 0.4818 | 0.7757 | 0.6672 | 0.7214 |
| 0.4027 | 9.0 | 12798 | 0.4871 | 0.7786 | 0.6722 | 0.7254 |
| 0.3926 | 10.0 | 14220 | 0.4919 | 0.7749 | 0.6932 | 0.7340 |
| 0.3824 | 11.0 | 15642 | 0.4890 | 0.7801 | 0.6823 | 0.7312 |
| 0.3718 | 12.0 | 17064 | 0.4981 | 0.7801 | 0.6970 | 0.7385 |
| 0.3629 | 13.0 | 18486 | 0.4989 | 0.7805 | 0.6968 | 0.7386 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.8.0
- Tokenizers 0.13.2