speaker-segmentation-fine-tuned-simsamu
This model is a fine-tuned version of pyannote/segmentation-3.0 on the diarizers-community/simsamu default dataset. It achieves the following results on the evaluation set:
- Loss: 0.2302
- Der: 0.0911
- False Alarm: 0.0236
- Missed Detection: 0.0413
- Confusion: 0.0262
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: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 5.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|
| 0.2179 | 1.0 | 111 | 0.2240 | 0.0964 | 0.0254 | 0.0470 | 0.0240 |
| 0.1678 | 2.0 | 222 | 0.2279 | 0.0943 | 0.0236 | 0.0447 | 0.0260 |
| 0.156 | 3.0 | 333 | 0.2327 | 0.0947 | 0.0222 | 0.0450 | 0.0274 |
| 0.1507 | 4.0 | 444 | 0.2301 | 0.0919 | 0.0237 | 0.0420 | 0.0262 |
| 0.1471 | 5.0 | 555 | 0.2302 | 0.0911 | 0.0236 | 0.0413 | 0.0262 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.19.1