indobert-base-p1-twitter-indonesia-sarcastic
This model is a fine-tuned version of indobenchmark/indobert-base-p1 on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.8125
- Accuracy: 0.8662
- F1: 0.7273
- Precision: 0.7385
- Recall: 0.7164
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: 1e-05
- train_batch_size: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 100.0
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|---|---|---|---|---|---|---|---|
| 0.5101 | 1.0 | 59 | 0.4221 | 0.7836 | 0.6081 | 0.5556 | 0.6716 |
| 0.3508 | 2.0 | 118 | 0.3479 | 0.8246 | 0.6713 | 0.6316 | 0.7164 |
| 0.221 | 3.0 | 177 | 0.3511 | 0.8582 | 0.6935 | 0.7544 | 0.6418 |
| 0.1157 | 4.0 | 236 | 0.4352 | 0.8396 | 0.6861 | 0.6714 | 0.7015 |
| 0.0453 | 5.0 | 295 | 0.6923 | 0.8582 | 0.7077 | 0.7302 | 0.6866 |
| 0.0192 | 6.0 | 354 | 0.7378 | 0.8694 | 0.7287 | 0.7581 | 0.7015 |
| 0.0159 | 7.0 | 413 | 0.8860 | 0.8545 | 0.6723 | 0.7692 | 0.5970 |
| 0.0165 | 8.0 | 472 | 0.8261 | 0.8694 | 0.7445 | 0.7286 | 0.7612 |
| 0.0175 | 9.0 | 531 | 0.8732 | 0.8731 | 0.7424 | 0.7538 | 0.7313 |
| 0.0062 | 10.0 | 590 | 0.9648 | 0.8657 | 0.7273 | 0.7385 | 0.7164 |
| 0.0003 | 11.0 | 649 | 1.0108 | 0.8619 | 0.7176 | 0.7344 | 0.7015 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.1+cu121
- Datasets 2.15.0
- Tokenizers 0.15.0