---
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: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# Whisper base AR - BA

This model is a fine-tuned version of [openai/whisper-base](https://huggingface.co/openai/whisper-base) on the quran-ayat-speech-to-text dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0036
- Wer: 0.0518
- Cer: 0.0212

## 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: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 2
- 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: 25
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch   | Step | Validation Loss | Wer    | Cer    |
|:-------------:|:-------:|:----:|:---------------:|:------:|:------:|
| 0.0051        | 1.0     | 282  | 0.0036          | 0.0472 | 0.0188 |
| 0.0033        | 2.0     | 564  | 0.0042          | 0.0547 | 0.0200 |
| 0.0018        | 3.0     | 846  | 0.0050          | 0.0599 | 0.0217 |
| 0.0012        | 4.0     | 1128 | 0.0054          | 0.0536 | 0.0195 |
| 0.0007        | 5.0     | 1410 | 0.0059          | 0.0599 | 0.0203 |
| 0.0006        | 6.0     | 1692 | 0.0069          | 0.0632 | 0.0238 |
| 0.0006        | 7.0     | 1974 | 0.0076          | 0.0673 | 0.0239 |
| 0.0004        | 8.0     | 2256 | 0.0076          | 0.0644 | 0.0235 |
| 0.0004        | 9.0     | 2538 | 0.0087          | 0.0603 | 0.0225 |
| 0.0003        | 10.0    | 2820 | 0.0076          | 0.0601 | 0.0220 |
| 0.0003        | 11.0    | 3102 | 0.0091          | 0.0615 | 0.0213 |
| 0.0002        | 12.0    | 3384 | 0.0094          | 0.0572 | 0.0211 |
| 0.0002        | 13.0    | 3666 | 0.0100          | 0.0686 | 0.0284 |
| 0.0001        | 14.0    | 3948 | 0.0102          | 0.0608 | 0.0221 |
| 0.0001        | 15.0    | 4230 | 0.0115          | 0.0552 | 0.0196 |
| 0.0001        | 16.0    | 4512 | 0.0119          | 0.0563 | 0.0208 |
| 0.0001        | 17.0    | 4794 | 0.0121          | 0.0614 | 0.0228 |
| 0.0           | 18.0    | 5076 | 0.0132          | 0.0559 | 0.0216 |
| 0.0           | 19.0    | 5358 | 0.0135          | 0.0568 | 0.0211 |
| 0.0           | 20.0    | 5640 | 0.0139          | 0.0559 | 0.0213 |
| 0.0           | 21.0    | 5922 | 0.0137          | 0.0552 | 0.0206 |
| 0.0           | 22.0    | 6204 | 0.0136          | 0.0541 | 0.0204 |
| 0.0           | 23.0    | 6486 | 0.0136          | 0.0538 | 0.0200 |
| 0.0           | 24.0    | 6768 | 0.0132          | 0.0608 | 0.0244 |
| 0.0           | 24.9130 | 7025 | 0.0136          | 0.0539 | 0.0201 |


### Framework versions

- Transformers 4.51.1
- Pytorch 2.5.1+cu124
- Datasets 3.5.0
- Tokenizers 0.21.0