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: 0.0746
- Wer: 0.2358
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 |
|---|---|---|---|---|
| 0.397 | 1.0 | 469 | 0.0686 | 0.2067 |
| 0.3526 | 2.0 | 938 | 0.0674 | 0.2345 |
| 0.3181 | 3.0 | 1407 | 0.0711 | 0.2225 |
| 0.2689 | 4.0 | 1876 | 0.0724 | 0.2273 |
| 0.2044 | 5.0 | 2345 | 0.0687 | 0.2366 |
| 0.184 | 6.0 | 2814 | 0.0676 | 0.2165 |
| 0.1404 | 7.0 | 3283 | 0.0673 | 0.2230 |
| 0.1461 | 8.0 | 3752 | 0.0657 | 0.2049 |
| 0.1269 | 9.0 | 4221 | 0.0662 | 0.2055 |
| 0.1089 | 10.0 | 4690 | 0.0668 | 0.2070 |
| 0.0835 | 11.0 | 5159 | 0.0639 | 0.2049 |
| 0.0693 | 12.0 | 5628 | 0.0656 | 0.2152 |
| 0.054 | 13.0 | 6097 | 0.0654 | 0.2160 |
| 0.0419 | 14.0 | 6566 | 0.0662 | 0.2029 |
| 0.0342 | 14.9685 | 7020 | 0.0668 | 0.2180 |
Framework versions
- Transformers 4.51.3
- Pytorch 2.6.0+cu124
- Datasets 3.6.0
- Tokenizers 0.21.1