metadata
license: mit
base_model: pyannote/segmentation-3.0
tags:
- speaker-diarization
- speaker-segmentation
- generated_from_trainer
datasets:
- ArtFair/diarizers_dataset_70-15-15
model-index:
- name: fine_tuned_segmentation-3.0_1e-3_128_pth
results: []
fine_tuned_segmentation-3.0_1e-3_128_pth
This model is a fine-tuned version of pyannote/segmentation-3.0 on the ArtFair/diarizers_dataset_70-15-15 default dataset. It achieves the following results on the evaluation set:
- Loss: 0.3620
- Der: 0.2625
- False Alarm: 0.1458
- Missed Detection: 0.0926
- Confusion: 0.0241
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: 128
- 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 |
|---|---|---|---|---|---|---|---|
| 0.426 | 1.0 | 233 | 0.3954 | 0.2915 | 0.1834 | 0.0807 | 0.0274 |
| 0.3974 | 2.0 | 466 | 0.3667 | 0.2668 | 0.1391 | 0.1032 | 0.0246 |
| 0.3772 | 3.0 | 699 | 0.3675 | 0.2672 | 0.1552 | 0.0874 | 0.0246 |
| 0.3618 | 4.0 | 932 | 0.3629 | 0.2641 | 0.1498 | 0.0899 | 0.0243 |
| 0.3622 | 5.0 | 1165 | 0.3620 | 0.2625 | 0.1458 | 0.0926 | 0.0241 |
Framework versions
- Transformers 4.37.2
- Pytorch 2.4.1+cu121
- Datasets 2.17.0
- Tokenizers 0.15.2