whisper-tiny-en
This model is a fine-tuned version of openai/whisper-tiny on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:
- Loss: 0.6638
- Wer Ortho: 0.3257
- Wer: 0.3264
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: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 50
- training_steps: 500
- mixed_precision_training: Native AMP
Training results
Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer |
---|---|---|---|---|---|
0.2837 | 3.57 | 100 | 0.4866 | 0.3356 | 0.3300 |
0.0286 | 7.14 | 200 | 0.5649 | 0.3294 | 0.3182 |
0.0021 | 10.71 | 300 | 0.6174 | 0.3276 | 0.3253 |
0.0009 | 14.29 | 400 | 0.6450 | 0.3257 | 0.3259 |
0.0006 | 17.86 | 500 | 0.6638 | 0.3257 | 0.3264 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.2.0a0+81ea7a4
- Datasets 2.18.0
- Tokenizers 0.15.2
- Downloads last month
- 13