|
--- |
|
language: |
|
- ar |
|
tags: |
|
- hf-asr-leaderboard |
|
- generated_from_trainer |
|
metrics: |
|
- wer |
|
model-index: |
|
- name: Whisper Small - Arabic language |
|
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 Small - Arabic language |
|
|
|
This model is a fine-tuned version of [MohammadJamalaldeen/whisper-small-with-google-fleurs-ar-4000_steps](https://huggingface.co/MohammadJamalaldeen/whisper-small-with-google-fleurs-ar-4000_steps). |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.5364 |
|
- Wer: 23.5875 |
|
|
|
## 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: 1e-05 |
|
- train_batch_size: 8 |
|
- eval_batch_size: 8 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: linear |
|
- lr_scheduler_warmup_steps: 500 |
|
- training_steps: 4000 |
|
- mixed_precision_training: Native AMP |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Wer | |
|
|:-------------:|:-----:|:----:|:---------------:|:-------:| |
|
| 0.0071 | 3.8 | 1000 | 0.4950 | 25.4875 | |
|
| 0.0003 | 7.6 | 2000 | 0.5170 | 23.8000 | |
|
| 0.0001 | 11.41 | 3000 | 0.5300 | 23.6125 | |
|
| 0.0001 | 15.21 | 4000 | 0.5364 | 23.5875 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.26.0 |
|
- Pytorch 1.13.1+cu116 |
|
- Datasets 2.7.0 |
|
- Tokenizers 0.13.2 |
|
|
