Whisper Small Mixed-Spanish
This model is a fine-tuned version of openai/whisper-small on the es datasets:
- mozilla-foundation/common_voice_17_0
- google/fleurs
- facebook/multilingual_librispeech
- facebook/voxpopuli
It achieves the following results on the evaluation set:
- Loss: 0.1809
- Wer: 8.6345
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: 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: 5000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.247 | 0.2 | 1000 | 0.2160 | 10.3975 |
| 0.1337 | 0.4 | 2000 | 0.2010 | 9.6749 |
| 0.1401 | 0.6 | 3000 | 0.1905 | 9.0946 |
| 0.1714 | 0.8 | 4000 | 0.1849 | 8.8550 |
| 0.1046 | 1.0 | 5000 | 0.1809 | 8.6345 |
Framework versions
- Transformers 4.42.0.dev0
- Pytorch 2.3.0+cu121
- Datasets 2.19.1
- Tokenizers 0.19.1