openai/whisper-tiny
This model is a fine-tuned version of openai/whisper-tiny on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.2034
- Wer: 5.3823
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: 64
- eval_batch_size: 32
- 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 |
|---|---|---|---|---|
| 0.0814 | 10.0 | 500 | 0.1915 | 6.6701 |
| 0.0045 | 20.0 | 1000 | 0.1816 | 5.5088 |
| 0.0016 | 30.01 | 1500 | 0.1924 | 5.5014 |
| 0.0009 | 40.01 | 2000 | 0.1959 | 5.5609 |
| 0.0006 | 51.0 | 2500 | 0.1989 | 5.4195 |
| 0.0005 | 61.0 | 3000 | 0.2014 | 5.4418 |
| 0.0004 | 71.01 | 3500 | 0.2030 | 5.3674 |
| 0.0004 | 81.01 | 4000 | 0.2034 | 5.3823 |
Framework versions
- Transformers 4.27.0.dev0
- Pytorch 1.13.1+cu117
- Datasets 2.9.1.dev0
- Tokenizers 0.13.2