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: 0.0955
- Wer: 0.2012
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 |
|---|---|---|---|---|
| 2.4056 | 1.0 | 313 | 0.0981 | 0.2034 |
| 2.413 | 2.0 | 626 | 0.0957 | 0.1978 |
| 1.9074 | 3.0 | 939 | 0.0975 | 0.2032 |
| 1.5982 | 4.0 | 1252 | 0.0961 | 0.1966 |
| 1.4477 | 5.0 | 1565 | 0.0950 | 0.2032 |
| 1.344 | 6.0 | 1878 | 0.0933 | 0.1951 |
| 1.1403 | 7.0 | 2191 | 0.0931 | 0.1944 |
| 1.0602 | 8.0 | 2504 | 0.0931 | 0.2007 |
| 1.0312 | 9.0 | 2817 | 0.0927 | 0.1947 |
| 0.9346 | 10.0 | 3130 | 0.0927 | 0.1927 |
| 0.8797 | 11.0 | 3443 | 0.0924 | 0.2006 |
| 0.8839 | 12.0 | 3756 | 0.0915 | 0.1964 |
| 0.8289 | 13.0 | 4069 | 0.0916 | 0.1929 |
| 0.7799 | 14.0 | 4382 | 0.0914 | 0.1973 |
| 0.7573 | 14.9536 | 4680 | 0.0914 | 0.1989 |
Framework versions
- Transformers 4.51.1
- Pytorch 2.5.1+cu124
- Datasets 3.5.0
- Tokenizers 0.21.0