speaker-segmentation-fine-tuned-callhome-eng-3
This model is a fine-tuned version of pyannote/segmentation-3.0 on the diarizers-community/callhome eng dataset. It achieves the following results on the evaluation set:
- Loss: 0.4652
- Der: 0.1821
- False Alarm: 0.0597
- Missed Detection: 0.0715
- Confusion: 0.0509
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: 64
- eval_batch_size: 64
- 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 |
---|---|---|---|---|---|---|---|
0.4563 | 1.0 | 181 | 0.4971 | 0.1973 | 0.0553 | 0.0802 | 0.0617 |
0.4053 | 2.0 | 362 | 0.4740 | 0.1899 | 0.0604 | 0.0749 | 0.0546 |
0.3833 | 3.0 | 543 | 0.4636 | 0.1854 | 0.0556 | 0.0766 | 0.0531 |
0.3738 | 4.0 | 724 | 0.4664 | 0.1830 | 0.0579 | 0.0733 | 0.0518 |
0.3596 | 5.0 | 905 | 0.4571 | 0.1800 | 0.0558 | 0.0748 | 0.0494 |
0.3533 | 6.0 | 1086 | 0.4671 | 0.1844 | 0.0629 | 0.0685 | 0.0529 |
0.3571 | 7.0 | 1267 | 0.4641 | 0.1820 | 0.0594 | 0.0711 | 0.0515 |
0.3496 | 8.0 | 1448 | 0.4641 | 0.1824 | 0.0596 | 0.0717 | 0.0511 |
0.3449 | 9.0 | 1629 | 0.4636 | 0.1819 | 0.0591 | 0.0718 | 0.0510 |
0.3415 | 10.0 | 1810 | 0.4652 | 0.1821 | 0.0597 | 0.0715 | 0.0509 |
Framework versions
- Transformers 4.40.1
- Pytorch 2.2.0+cu121
- Datasets 2.17.0
- Tokenizers 0.19.1
- Downloads last month
- 2
Model tree for tgrhn/speaker-segmentation-fine-tuned-callhome-eng-3
Base model
pyannote/segmentation-3.0