arabic-sentiment-finetuned
This model is a fine-tuned version of Walid-Ahmed/arabic-sentiment-model on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3205
- Accuracy: 0.8845
- F1: 0.8844
- Precision: 0.8850
- Recall: 0.8838
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: 2e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 0.1
- num_epochs: 5
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|---|---|---|---|---|---|---|---|
| 0.3448 | 1.0 | 1420 | 0.3155 | 0.8769 | 0.8766 | 0.8789 | 0.8743 |
| 0.2809 | 2.0 | 2840 | 0.3196 | 0.8849 | 0.8848 | 0.8854 | 0.8842 |
| 0.1962 | 3.0 | 4260 | 0.3735 | 0.8835 | 0.8843 | 0.8776 | 0.8912 |
| 0.1805 | 4.0 | 5680 | 0.4175 | 0.8776 | 0.8789 | 0.8700 | 0.8880 |
| 0.1660 | 5.0 | 7100 | 0.4682 | 0.8761 | 0.8771 | 0.8696 | 0.8849 |
Framework versions
- Transformers 5.0.0
- Pytorch 2.10.0+cu128
- Datasets 4.0.0
- Tokenizers 0.22.2