distilbert_sa_GLUE_Experiment_logit_kd_pretrain_qqp
This model is a fine-tuned version of gokuls/distilbert_sa_pre-training-complete on the GLUE QQP dataset. It achieves the following results on the evaluation set:
- Loss: 0.5449
- Accuracy: 0.6632
- F1: 0.1647
- Combined Score: 0.4139
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 |
|---|---|---|---|---|---|---|
| 0.6004 | 1.0 | 1422 | 0.5643 | 0.6623 | 0.1630 | 0.4126 |
| 0.5393 | 2.0 | 2844 | 0.5498 | 0.6538 | 0.1199 | 0.3869 |
| 0.5157 | 3.0 | 4266 | 0.5449 | 0.6632 | 0.1647 | 0.4139 |
| 0.5007 | 4.0 | 5688 | 0.5512 | 0.6848 | 0.2663 | 0.4755 |
| 0.4914 | 5.0 | 7110 | 0.5501 | 0.6665 | 0.1817 | 0.4241 |
| 0.4847 | 6.0 | 8532 | 0.5475 | 0.6816 | 0.2517 | 0.4667 |
| 0.4803 | 7.0 | 9954 | 0.5478 | 0.6768 | 0.2301 | 0.4535 |
| 0.4768 | 8.0 | 11376 | 0.5488 | 0.6839 | 0.2610 | 0.4724 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.14.0a0+410ce96
- Datasets 2.9.0
- Tokenizers 0.13.2