distilbert_sa_GLUE_Experiment_logit_kd_data_aug_mrpc_96
This model is a fine-tuned version of distilbert-base-uncased on the GLUE MRPC dataset. It achieves the following results on the evaluation set:
- Loss: 0.3968
- Accuracy: 1.0
- F1: 1.0
- Combined Score: 1.0
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 |
|---|---|---|---|---|---|---|
| 0.4718 | 1.0 | 980 | 0.4083 | 0.875 | 0.8998 | 0.8874 |
| 0.4239 | 2.0 | 1960 | 0.3986 | 0.9902 | 0.9928 | 0.9915 |
| 0.4182 | 3.0 | 2940 | 0.3981 | 0.9877 | 0.9910 | 0.9894 |
| 0.4155 | 4.0 | 3920 | 0.3975 | 0.9828 | 0.9873 | 0.9851 |
| 0.4146 | 5.0 | 4900 | 0.3969 | 0.9902 | 0.9928 | 0.9915 |
| 0.4136 | 6.0 | 5880 | 0.3970 | 0.9853 | 0.9891 | 0.9872 |
| 0.4131 | 7.0 | 6860 | 0.3970 | 0.9926 | 0.9946 | 0.9936 |
| 0.4127 | 8.0 | 7840 | 0.3968 | 1.0 | 1.0 | 1.0 |
| 0.4123 | 9.0 | 8820 | 0.3970 | 1.0 | 1.0 | 1.0 |
| 0.412 | 10.0 | 9800 | 0.3969 | 1.0 | 1.0 | 1.0 |
| 0.4119 | 11.0 | 10780 | 0.3969 | 0.9926 | 0.9946 | 0.9936 |
| 0.4116 | 12.0 | 11760 | 0.3968 | 0.9902 | 0.9928 | 0.9915 |
| 0.4116 | 13.0 | 12740 | 0.3968 | 0.9951 | 0.9964 | 0.9958 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2