tiny-mlm-glue-qnli-target-glue-mrpc
This model is a fine-tuned version of muhtasham/tiny-mlm-glue-qnli on the None dataset. It achieves the following results on the evaluation set:
- Loss: 1.3118
- Accuracy: 0.7353
- F1: 0.8176
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: 3e-05
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- num_epochs: 200
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|---|---|---|---|---|---|
| 0.5845 | 4.35 | 500 | 0.5424 | 0.7402 | 0.8290 |
| 0.4395 | 8.7 | 1000 | 0.5503 | 0.7525 | 0.8374 |
| 0.2883 | 13.04 | 1500 | 0.6404 | 0.7475 | 0.8280 |
| 0.1828 | 17.39 | 2000 | 0.7736 | 0.7574 | 0.8406 |
| 0.1141 | 21.74 | 2500 | 1.0144 | 0.7255 | 0.8056 |
| 0.0816 | 26.09 | 3000 | 1.1432 | 0.7328 | 0.8180 |
| 0.0616 | 30.43 | 3500 | 1.3118 | 0.7353 | 0.8176 |
Framework versions
- Transformers 4.26.0.dev0
- Pytorch 1.13.0+cu116
- Datasets 2.8.1.dev0
- Tokenizers 0.13.2