Whisper-Small-Inbrowser-Proctor
This model is a fine-tuned version of openai/whisper-small on the Inbrowser Procotor Dataset dataset. It achieves the following results on the evaluation set:
- Loss: 0.3099
- Wer: 17.0755
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: 5e-06
- train_batch_size: 8
- 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: 25
- training_steps: 250
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Wer |
|---|---|---|---|---|
| 0.3461 | 0.4545 | 25 | 0.4545 | 26.0433 |
| 0.1902 | 0.9091 | 50 | 0.3309 | 17.4419 |
| 0.1184 | 1.3636 | 75 | 0.3120 | 14.6543 |
| 0.0944 | 1.8182 | 100 | 0.3066 | 16.7251 |
| 0.0632 | 2.2727 | 125 | 0.3046 | 14.8455 |
| 0.0688 | 2.7273 | 150 | 0.3060 | 14.8933 |
| 0.0479 | 3.1818 | 175 | 0.3063 | 17.1074 |
| 0.0515 | 3.6364 | 200 | 0.3081 | 15.4986 |
| 0.0296 | 4.0909 | 225 | 0.3096 | 17.2507 |
| 0.0348 | 4.5455 | 250 | 0.3099 | 17.0755 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.0
- Tokenizers 0.19.1