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: 0.3412
- Der: 0.1116
- False Alarm: 0.0194
- Missed Detection: 0.0267
- Confusion: 0.0655
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: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine
- num_epochs: 5
Training results
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|
| 0.3963 | 1.0 | 755 | 0.4401 | 0.1468 | 0.0203 | 0.0353 | 0.0912 |
| 0.3814 | 2.0 | 1510 | 0.3740 | 0.1222 | 0.0185 | 0.0310 | 0.0727 |
| 0.3112 | 3.0 | 2265 | 0.3522 | 0.1192 | 0.0209 | 0.0263 | 0.0720 |
| 0.3186 | 4.0 | 3020 | 0.3417 | 0.1127 | 0.0188 | 0.0275 | 0.0664 |
| 0.3093 | 5.0 | 3775 | 0.3412 | 0.1116 | 0.0194 | 0.0267 | 0.0655 |
Framework versions
- Transformers 4.48.3
- Pytorch 2.5.1+cu124
- Datasets 3.3.1
- Tokenizers 0.21.0
- Downloads last month
- 11
Inference Providers
NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API:
The model has no pipeline_tag.
Model tree for nairaxo/diarization-model-ary
Base model
pyannote/speaker-diarization-3.1