Whisper Small for Even Pakendorf
This model is a fine-tuned version of openai/whisper-small on the Even Speech Pakendorf dataset. It achieves the following results on the evaluation set:
- Loss: 1.3086
- Wer: 67.9879
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 |
|---|---|---|---|---|
| 0.6619 | 2.6954 | 1000 | 0.9961 | 86.2737 |
| 0.0989 | 5.3908 | 2000 | 1.1101 | 68.8140 |
| 0.0114 | 8.0863 | 3000 | 1.2432 | 68.4645 |
| 0.0049 | 10.7817 | 4000 | 1.3086 | 67.9879 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.1+cu118
- Datasets 2.20.0
- Tokenizers 0.19.1
- Downloads last month
- 5