metadata
library_name: transformers
language:
- jpn
license: mit
base_model: pyannote/speaker-diarization-3.1
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/speaker-diarization-3.1 on the diarizers-community/callhome dataset. It achieves the following results on the evaluation set:
- Loss: 0.4370
- Model Preparation Time: 0.0038
- Der: 0.1404
- False Alarm: 0.0234
- Missed Detection: 0.0272
- Confusion: 0.0898
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 |
|---|---|---|---|---|---|---|---|---|
| 0.4567 | 1.0 | 194 | 0.4782 | 0.0038 | 0.1572 | 0.0219 | 0.0351 | 0.1003 |
| 0.384 | 2.0 | 388 | 0.4576 | 0.0038 | 0.1483 | 0.0221 | 0.0295 | 0.0968 |
| 0.3582 | 3.0 | 582 | 0.4375 | 0.0038 | 0.1402 | 0.0230 | 0.0284 | 0.0888 |
| 0.357 | 4.0 | 776 | 0.4406 | 0.0038 | 0.1413 | 0.0229 | 0.0277 | 0.0907 |
| 0.3406 | 5.0 | 970 | 0.4370 | 0.0038 | 0.1404 | 0.0234 | 0.0272 | 0.0898 |
Framework versions
- Transformers 4.44.2
- Pytorch 2.5.0+cu121
- Datasets 3.1.0
- Tokenizers 0.19.1