speaker-segmentation-fine-tuned-callhome-eng-forproject
This model is a fine-tuned version of pyannote/speaker-diarization-3.1 on the diarizers-community/callhome dataset. It achieves the following results on the evaluation set:
- Loss: 0.4600
- Model Preparation Time: 0.0051
- Der: 0.1818
- False Alarm: 0.0578
- Missed Detection: 0.0721
- Confusion: 0.0518
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 | Model Preparation Time | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|---|
| 0.392 | 1.0 | 362 | 0.4730 | 0.0051 | 0.1926 | 0.0622 | 0.0736 | 0.0568 |
| 0.4053 | 2.0 | 724 | 0.4586 | 0.0051 | 0.1838 | 0.0625 | 0.0704 | 0.0509 |
| 0.3865 | 3.0 | 1086 | 0.4537 | 0.0051 | 0.1811 | 0.0574 | 0.0723 | 0.0514 |
| 0.3571 | 4.0 | 1448 | 0.4570 | 0.0051 | 0.1805 | 0.0551 | 0.0740 | 0.0514 |
| 0.3409 | 5.0 | 1810 | 0.4600 | 0.0051 | 0.1818 | 0.0578 | 0.0721 | 0.0518 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.4.1+cu121
- Datasets 3.0.1
- Tokenizers 0.19.1
- Downloads last month
- 14
Model tree for JSWOOK/speaker-segmentation-fine-tuned-callhome-eng
Base model
pyannote/speaker-diarization-3.1