Whisper Small Hi - Satish Lokkoju
This model is a fine-tuned version of openai/whisper-small on the Common Voice 13.0 dataset. It achieves the following results on the evaluation set:
- Loss: 0.4093
- Wer: 31.8357
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: 500
- training_steps: 4000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.0824 | 2.36 | 1000 | 0.2878 | 34.6474 |
| 0.0214 | 4.73 | 2000 | 0.3354 | 33.0407 |
| 0.0013 | 7.09 | 3000 | 0.3846 | 32.0883 |
| 0.0006 | 9.46 | 4000 | 0.4093 | 31.8357 |
Framework versions
- Transformers 4.27.4
- Pytorch 2.0.0+cu117
- Datasets 2.11.0
- Tokenizers 0.13.3
- Downloads last month
- 9
Dataset used to train lsatish/whisper-small-hi
Evaluation results
- Wer on Common Voice 13.0test set self-reported31.836