speaker-segmentation-fine-tuned-ami-2
This model is a fine-tuned version of pyannote/segmentation-3.0 on the diarizers-community/ami ihm dataset. It achieves the following results on the evaluation set:
- Loss: 0.3764
- Der: 0.1401
- False Alarm: 0.0503
- Missed Detection: 0.0575
- Confusion: 0.0323
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: 10.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|
| 0.4149 | 1.0 | 1427 | 0.3607 | 0.1407 | 0.0492 | 0.0593 | 0.0323 |
| 0.3915 | 2.0 | 2854 | 0.3684 | 0.1422 | 0.0460 | 0.0621 | 0.0340 |
| 0.3748 | 3.0 | 4281 | 0.3730 | 0.1419 | 0.0530 | 0.0570 | 0.0318 |
| 0.3778 | 4.0 | 5708 | 0.3649 | 0.1409 | 0.0472 | 0.0611 | 0.0326 |
| 0.3565 | 5.0 | 7135 | 0.3723 | 0.1415 | 0.0501 | 0.0591 | 0.0324 |
| 0.3566 | 6.0 | 8562 | 0.3740 | 0.1406 | 0.0499 | 0.0584 | 0.0323 |
| 0.3534 | 7.0 | 9989 | 0.3736 | 0.1399 | 0.0493 | 0.0581 | 0.0325 |
| 0.3418 | 8.0 | 11416 | 0.3744 | 0.1397 | 0.0500 | 0.0577 | 0.0321 |
| 0.3388 | 9.0 | 12843 | 0.3777 | 0.1403 | 0.0505 | 0.0574 | 0.0324 |
| 0.346 | 10.0 | 14270 | 0.3764 | 0.1401 | 0.0503 | 0.0575 | 0.0323 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.19.1