whisper-small-finetuned-400-standard-A-epochs-10
This model is a fine-tuned version of openai/whisper-small on the horoscope_standard_a_400_19_20_5_03 dataset. It achieves the following results on the evaluation set:
- Loss: 0.5064
- Wer: 27.4148
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: 8
- eval_batch_size: 4
- seed: 42
- optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| No log | 0 | 0 | 0.6890 | 41.0325 |
| 0.1722 | 1.0 | 50 | 0.4912 | 30.1563 |
| 0.0365 | 2.0 | 100 | 0.4825 | 28.5678 |
| 0.0134 | 3.0 | 150 | 0.4958 | 27.7479 |
| 0.0065 | 4.0 | 200 | 0.5046 | 28.3500 |
| 0.0042 | 5.0 | 250 | 0.5026 | 27.6326 |
| 0.0027 | 6.0 | 300 | 0.5018 | 27.5045 |
| 0.0019 | 7.0 | 350 | 0.5035 | 27.4789 |
| 0.0017 | 8.0 | 400 | 0.5048 | 27.3123 |
| 0.0015 | 9.0 | 450 | 0.5060 | 27.4276 |
| 0.0015 | 10.0 | 500 | 0.5064 | 27.4148 |
Framework versions
- Transformers 4.49.0
- Pytorch 2.5.1+cu121
- Datasets 3.4.0
- Tokenizers 0.21.0