metadata
library_name: transformers
language:
- ar
license: apache-2.0
base_model: openai/whisper-base
tags:
- generated_from_trainer
metrics:
- wer
model-index:
- name: Whisper base AR - BA
results: []
Whisper base AR - BA
This model is a fine-tuned version of openai/whisper-base on the quran-ayat-speech-to-text dataset. It achieves the following results on the evaluation set:
- Loss: 0.1829
- Wer: 0.4490
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 32
- optimizer: Use OptimizerNames.ADAMW_TORCH 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: 500
- num_epochs: 15
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 12.086 | 1.0 | 469 | 0.1385 | 0.5869 |
| 3.7427 | 2.0 | 938 | 0.1307 | 0.5248 |
| 1.8383 | 3.0 | 1407 | 0.1461 | 0.4864 |
| 1.1846 | 4.0 | 1876 | 0.1576 | 0.4844 |
| 0.7395 | 5.0 | 2345 | 0.1697 | 0.5059 |
| 0.4648 | 6.0 | 2814 | 0.1640 | 0.4828 |
| 0.3456 | 7.0 | 3283 | 0.1759 | 0.4794 |
| 0.3017 | 8.0 | 3752 | 0.1790 | 0.4945 |
| 0.2196 | 9.0 | 4221 | 0.1723 | 0.4749 |
| 0.1881 | 10.0 | 4690 | 0.1639 | 0.4663 |
| 0.1507 | 11.0 | 5159 | 0.1688 | 0.4307 |
| 0.1221 | 12.0 | 5628 | 0.1595 | 0.4555 |
| 0.1179 | 13.0 | 6097 | 0.1617 | 0.4642 |
| 0.0914 | 14.0 | 6566 | 0.1611 | 0.4636 |
| 0.065 | 14.9685 | 7020 | 0.1634 | 0.4623 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1