Whisper Tiny Ar Tashkeel - AzeemX
This model is a fine-tuned version of tarteel-ai/whisper-tiny-ar-quran on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.1126
- Wer: 18.6851
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: 16
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 500
- training_steps: 7000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.2248 | 1.0320 | 1000 | 0.2138 | 34.5276 |
| 0.1345 | 2.0640 | 2000 | 0.1635 | 27.2003 |
| 0.0967 | 3.0960 | 3000 | 0.1408 | 23.9162 |
| 0.0678 | 4.1280 | 4000 | 0.1282 | 21.5152 |
| 0.0537 | 5.1600 | 5000 | 0.1191 | 19.9726 |
| 0.0501 | 6.1920 | 6000 | 0.1145 | 19.0116 |
| 0.0416 | 7.2239 | 7000 | 0.1126 | 18.6851 |
Framework versions
- Transformers 4.45.1
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.20.0