<!-- 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. -->
distilbert_sa_GLUE_Experiment_logit_kd_data_aug_mrpc
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: 0.3959
- Accuracy: 0.8946
- F1: 0.9165
- Combined Score: 0.9056
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 |
---|---|---|---|---|---|---|
0.4396 | 1.0 | 980 | 0.4005 | 0.9975 | 0.9982 | 0.9979 |
0.4146 | 2.0 | 1960 | 0.3981 | 0.9853 | 0.9892 | 0.9872 |
0.413 | 3.0 | 2940 | 0.3971 | 1.0 | 1.0 | 1.0 |
0.4119 | 4.0 | 3920 | 0.3968 | 0.9902 | 0.9928 | 0.9915 |
0.4114 | 5.0 | 4900 | 0.3966 | 1.0 | 1.0 | 1.0 |
0.4112 | 6.0 | 5880 | 0.3966 | 0.9951 | 0.9964 | 0.9958 |
0.4108 | 7.0 | 6860 | 0.3966 | 1.0 | 1.0 | 1.0 |
0.4107 | 8.0 | 7840 | 0.3962 | 0.9877 | 0.9910 | 0.9894 |
0.4105 | 9.0 | 8820 | 0.3962 | 0.9902 | 0.9928 | 0.9915 |
0.4104 | 10.0 | 9800 | 0.3967 | 0.9020 | 0.9228 | 0.9124 |
0.4102 | 11.0 | 10780 | 0.3964 | 0.8971 | 0.9186 | 0.9078 |
0.4102 | 12.0 | 11760 | 0.3963 | 0.9975 | 0.9982 | 0.9979 |
0.4103 | 13.0 | 12740 | 0.3962 | 0.8431 | 0.8704 | 0.8568 |
0.4102 | 14.0 | 13720 | 0.3967 | 0.7966 | 0.8253 | 0.8109 |
0.4101 | 15.0 | 14700 | 0.3962 | 0.8971 | 0.9186 | 0.9078 |
0.4101 | 16.0 | 15680 | 0.3959 | 0.8946 | 0.9165 | 0.9056 |
0.41 | 17.0 | 16660 | 0.3963 | 0.8848 | 0.9080 | 0.8964 |
0.41 | 18.0 | 17640 | 0.3960 | 0.8676 | 0.8929 | 0.8803 |
0.41 | 19.0 | 18620 | 0.3962 | 0.8186 | 0.8471 | 0.8329 |
0.4101 | 20.0 | 19600 | 0.3959 | 0.8848 | 0.9080 | 0.8964 |
0.41 | 21.0 | 20580 | 0.3959 | 0.8799 | 0.9037 | 0.8918 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2