speaker-segmentation-fine-tuned-simsamu-2
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: 0.2428
- Der: 0.0861
- False Alarm: 0.0245
- Missed Detection: 0.0384
- Confusion: 0.0232
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: 10.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|
| 0.2179 | 1.0 | 111 | 0.2259 | 0.0951 | 0.0239 | 0.0486 | 0.0227 |
| 0.1694 | 2.0 | 222 | 0.2379 | 0.0930 | 0.0230 | 0.0466 | 0.0234 |
| 0.1559 | 3.0 | 333 | 0.2305 | 0.0898 | 0.0223 | 0.0431 | 0.0244 |
| 0.149 | 4.0 | 444 | 0.2323 | 0.0893 | 0.0246 | 0.0398 | 0.0249 |
| 0.1416 | 5.0 | 555 | 0.2351 | 0.0884 | 0.0243 | 0.0399 | 0.0243 |
| 0.1369 | 6.0 | 666 | 0.2458 | 0.0904 | 0.0266 | 0.0370 | 0.0268 |
| 0.1367 | 7.0 | 777 | 0.2410 | 0.0882 | 0.0204 | 0.0434 | 0.0244 |
| 0.1306 | 8.0 | 888 | 0.2400 | 0.0866 | 0.0240 | 0.0393 | 0.0234 |
| 0.1301 | 9.0 | 999 | 0.2422 | 0.0860 | 0.0243 | 0.0387 | 0.0230 |
| 0.1276 | 10.0 | 1110 | 0.2428 | 0.0861 | 0.0245 | 0.0384 | 0.0232 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.19.1