Whisper Small Malayalam - Arjun Shaji
This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.7363
- Wer: 81.6092
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: 16
- 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: 1000
- training_steps: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0433 | 18.5185 | 500 | 0.5265 | 94.7126 |
| 0.0144 | 37.0370 | 1000 | 0.5352 | 89.1954 |
| 0.0057 | 55.5556 | 1500 | 0.5989 | 87.5862 |
| 0.0004 | 74.0741 | 2000 | 0.6575 | 82.0690 |
| 0.0 | 92.5926 | 2500 | 0.6616 | 81.6092 |
| 0.0 | 111.1111 | 3000 | 0.6911 | 81.3793 |
| 0.0 | 129.6296 | 3500 | 0.7097 | 81.3793 |
| 0.0 | 148.1481 | 4000 | 0.7232 | 81.3793 |
| 0.0 | 166.6667 | 4500 | 0.7327 | 81.3793 |
| 0.0 | 185.1852 | 5000 | 0.7363 | 81.6092 |
Framework versions
- Transformers 4.41.0
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1