speaker-segmentation-fine-tuned-callhome-spa
This model is a fine-tuned version of openai/whisper-small on the diarizers-community/callhome dataset. It achieves the following results on the evaluation set:
- Loss: 0.5174
- Der: 0.1732
- False Alarm: 0.0744
- Missed Detection: 0.0663
- Confusion: 0.0325
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: 0.001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|
| 0.6346 | 1.0 | 382 | 0.5316 | 0.1789 | 0.0670 | 0.0750 | 0.0369 |
| 0.6117 | 2.0 | 764 | 0.5156 | 0.1724 | 0.0648 | 0.0766 | 0.0311 |
| 0.6017 | 3.0 | 1146 | 0.5150 | 0.1747 | 0.0737 | 0.0680 | 0.0331 |
| 0.6173 | 4.0 | 1528 | 0.5162 | 0.1737 | 0.0748 | 0.0663 | 0.0326 |
| 0.5914 | 5.0 | 1910 | 0.5174 | 0.1732 | 0.0744 | 0.0663 | 0.0325 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.1+cu121
- Datasets 2.19.0
- Tokenizers 0.19.1