metadata
license: mit
base_model: pyannote/segmentation-3.0
tags:
- speaker-diarization
- speaker-segmentation
- generated_from_trainer
datasets:
- diarizers-community/callhome
model-index:
- name: speaker-segmentation-fine-tuned-callhome-jpn
results: []
speaker-segmentation-fine-tuned-callhome-jpn
This model is a fine-tuned version of pyannote/segmentation-3.0 on the diarizers-community/callhome jpn dataset. It achieves the following results on the evaluation set:
- Loss: 0.5146
- Der: 0.1869
- False Alarm: 0.0933
- Missed Detection: 0.0709
- Confusion: 0.0227
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: 5.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|
| 0.6043 | 1.0 | 340 | 0.5075 | 0.1789 | 0.0682 | 0.0789 | 0.0318 |
| 0.5766 | 2.0 | 680 | 0.5207 | 0.1951 | 0.1012 | 0.0708 | 0.0230 |
| 0.5345 | 3.0 | 1020 | 0.5011 | 0.1798 | 0.0852 | 0.0716 | 0.0231 |
| 0.518 | 4.0 | 1360 | 0.5344 | 0.1934 | 0.1009 | 0.0700 | 0.0225 |
| 0.5147 | 5.0 | 1700 | 0.5146 | 0.1869 | 0.0933 | 0.0709 | 0.0227 |
Framework versions
- Transformers 4.40.0
- Pytorch 2.2.2+cu121
- Datasets 2.18.0
- Tokenizers 0.19.1