distilled-indobert-classification
This model is a fine-tuned version of distilbert-base-uncased on the indonlu dataset. It achieves the following results on the evaluation set:
- Loss: 0.6015
- Accuracy: 0.9016
- F1: 0.9015
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: 6e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 33
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 |
|---|---|---|---|---|---|
| 1.0427 | 1.0 | 688 | 0.6306 | 0.8683 | 0.8684 |
| 0.5332 | 2.0 | 1376 | 0.5621 | 0.8794 | 0.8779 |
| 0.3021 | 3.0 | 2064 | 0.6785 | 0.8905 | 0.8896 |
| 0.1851 | 4.0 | 2752 | 0.6085 | 0.8968 | 0.8959 |
| 0.1152 | 5.0 | 3440 | 0.6015 | 0.9016 | 0.9015 |
Framework versions
- Transformers 4.18.0
- Pytorch 1.10.0+cu111
- Datasets 2.0.0
- Tokenizers 0.11.6