<!-- 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_qqp
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: 0.6308
- Accuracy: 0.6473
- F1: 0.0880
- Combined Score: 0.3676
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 |
---|---|---|---|---|---|---|
0.7821 | 1.0 | 1422 | 0.7485 | 0.6318 | 0.0 | 0.3159 |
0.7105 | 2.0 | 2844 | 0.7038 | 0.6364 | 0.0261 | 0.3312 |
0.6654 | 3.0 | 4266 | 0.6862 | 0.6351 | 0.0188 | 0.3269 |
0.6284 | 4.0 | 5688 | 0.6610 | 0.6453 | 0.0779 | 0.3616 |
0.5969 | 5.0 | 7110 | 0.6479 | 0.6416 | 0.0554 | 0.3485 |
0.5712 | 6.0 | 8532 | 0.6457 | 0.6404 | 0.0497 | 0.3450 |
0.5513 | 7.0 | 9954 | 0.6308 | 0.6473 | 0.0880 | 0.3676 |
0.5349 | 8.0 | 11376 | 0.6351 | 0.6503 | 0.1037 | 0.3770 |
0.5222 | 9.0 | 12798 | 0.6383 | 0.6719 | 0.2134 | 0.4427 |
0.5124 | 10.0 | 14220 | 0.6392 | 0.6685 | 0.1991 | 0.4338 |
0.5044 | 11.0 | 15642 | 0.6379 | 0.6615 | 0.1631 | 0.4123 |
0.4978 | 12.0 | 17064 | 0.6363 | 0.6637 | 0.1750 | 0.4194 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2