Whisper Tunisien
This model is a fine-tuned version of openai/whisper-small on the Tunisian_dataset_STT-TTS15s_filtred1.0 dataset. It achieves the following results on the evaluation set:
- Loss: 3.0846
- Wer: 109.9932
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-07
- 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: 3000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 1.6852 | 3.8760 | 500 | 3.5237 | 148.8184 |
| 1.2494 | 7.7519 | 1000 | 3.2200 | 121.6070 |
| 1.2202 | 11.6279 | 1500 | 3.1493 | 125.3883 |
| 1.0905 | 15.5039 | 2000 | 3.1099 | 113.9095 |
| 1.0606 | 19.3798 | 2500 | 3.0905 | 110.1958 |
| 1.0858 | 23.2558 | 3000 | 3.0846 | 109.9932 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
- Downloads last month
- 3
Model tree for Arbi-Houssem/TunLangModel_test1.10
Base model
openai/whisper-smallDataset used to train Arbi-Houssem/TunLangModel_test1.10
Evaluation results
- Wer on Tunisian_dataset_STT-TTS15s_filtred1.0self-reported109.993