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: 0.1423
- Wer: 0.4671
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 |
|---|---|---|---|---|
| 50.046 | 1.0 | 469 | 0.1557 | 0.5936 |
| 14.5095 | 2.0 | 938 | 0.1451 | 0.4789 |
| 7.9218 | 3.0 | 1407 | 0.1457 | 0.4676 |
| 5.2637 | 4.0 | 1876 | 0.1451 | 0.4367 |
| 4.0077 | 5.0 | 2345 | 0.1449 | 0.4478 |
| 2.8467 | 6.0 | 2814 | 0.1424 | 0.4400 |
| 2.3483 | 7.0 | 3283 | 0.1336 | 0.4219 |
| 1.9948 | 8.0 | 3752 | 0.1343 | 0.4472 |
| 1.7564 | 9.0 | 4221 | 0.1313 | 0.4217 |
| 1.6001 | 10.0 | 4690 | 0.1324 | 0.4345 |
| 1.4008 | 11.0 | 5159 | 0.1300 | 0.4190 |
| 1.283 | 12.0 | 5628 | 0.1300 | 0.4360 |
| 1.1674 | 13.0 | 6097 | 0.1299 | 0.4260 |
| 1.0709 | 14.0 | 6566 | 0.1293 | 0.4172 |
| 0.9897 | 14.9685 | 7020 | 0.1288 | 0.4413 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1