MiniLMv2-L12-H384-distilled-finetuned-clinc
This model is a fine-tuned version of nreimers/MiniLMv2-L12-H384-distilled-from-RoBERTa-Large on the clinc_oos dataset. It achieves the following results on the evaluation set:
- Loss: 0.3058
- Accuracy: 0.9529
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: 0.0001
- train_batch_size: 64
- eval_batch_size: 64
- seed: 33
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.9908 | 1.0 | 239 | 1.6816 | 0.3910 |
| 1.5212 | 2.0 | 478 | 1.2365 | 0.7697 |
| 1.129 | 3.0 | 717 | 0.9209 | 0.8706 |
| 0.8462 | 4.0 | 956 | 0.6978 | 0.9152 |
| 0.6497 | 5.0 | 1195 | 0.5499 | 0.9342 |
| 0.5124 | 6.0 | 1434 | 0.4447 | 0.9445 |
| 0.4196 | 7.0 | 1673 | 0.3797 | 0.9455 |
| 0.3587 | 8.0 | 1912 | 0.3358 | 0.95 |
| 0.3228 | 9.0 | 2151 | 0.3133 | 0.9513 |
| 0.3052 | 10.0 | 2390 | 0.3058 | 0.9529 |
Framework versions
- Transformers 4.17.0
- Pytorch 1.10.2+cu113
- Datasets 1.18.4
- Tokenizers 0.13.0