metadata
license: apache-2.0
base_model: openai/whisper-medium
tags:
- generated_from_trainer
model-index:
- name: Whisper-medium-Arabic-phoneme-10
results: []
Whisper-medium-Arabic-phoneme-10
This model is a fine-tuned version of openai/whisper-medium on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.2343
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: 0.001
- 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: 50
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 2.815 | 1.0 | 41 | 0.3574 |
| 0.1191 | 2.0 | 82 | 0.2477 |
| 0.0683 | 3.0 | 123 | 0.2215 |
| 0.0374 | 4.0 | 164 | 0.2570 |
| 0.0233 | 5.0 | 205 | 0.2492 |
| 0.0139 | 6.0 | 246 | 0.2441 |
| 0.0054 | 7.0 | 287 | 0.2294 |
| 0.0015 | 8.0 | 328 | 0.2363 |
| 0.0008 | 9.0 | 369 | 0.2335 |
| 0.0006 | 10.0 | 410 | 0.2343 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0