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: 0.0928
- Wer: 0.2043
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 |
|---|---|---|---|---|
| 1.2944 | 1.0 | 313 | 0.0886 | 0.1967 |
| 1.2819 | 2.0 | 626 | 0.0902 | 0.1923 |
| 1.2752 | 3.0 | 939 | 0.0902 | 0.1986 |
| 1.1425 | 4.0 | 1252 | 0.0915 | 0.1989 |
| 1.0812 | 5.0 | 1565 | 0.0900 | 0.1914 |
| 0.9708 | 6.0 | 1878 | 0.0900 | 0.1916 |
| 0.9029 | 7.0 | 2191 | 0.0891 | 0.1985 |
| 0.8248 | 8.0 | 2504 | 0.0896 | 0.1916 |
| 0.7778 | 9.0 | 2817 | 0.0897 | 0.1941 |
| 0.7485 | 10.0 | 3130 | 0.0890 | 0.1944 |
| 0.7219 | 11.0 | 3443 | 0.0883 | 0.1961 |
| 0.6584 | 12.0 | 3756 | 0.0889 | 0.1948 |
| 0.6516 | 13.0 | 4069 | 0.0883 | 0.1951 |
| 0.6233 | 14.0 | 4382 | 0.0882 | 0.1942 |
| 0.6017 | 14.9536 | 4680 | 0.0883 | 0.1957 |
Framework versions
- Transformers 4.51.1
- Pytorch 2.5.1+cu124
- Datasets 3.5.0
- Tokenizers 0.21.0