metadata
license: mit
base_model: pyannote/segmentation-3.0
tags:
- speaker-diarization
- speaker-segmentation
- generated_from_trainer
datasets:
- sparkleai/Diarizers_dataset_test
model-index:
- name: Diarizers_finetuned_model_test
results: []
Diarizers_finetuned_model_test
This model is a fine-tuned version of pyannote/segmentation-3.0 on the sparkleai/Diarizers_dataset_test default dataset. It achieves the following results on the evaluation set:
- Loss: 0.7692
- Der: 0.2484
- False Alarm: 0.0450
- Missed Detection: 0.0910
- Confusion: 0.1124
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 |
|---|---|---|---|---|---|---|---|
| No log | 1.0 | 6 | 0.8050 | 0.2813 | 0.0708 | 0.1119 | 0.0986 |
| No log | 2.0 | 12 | 0.7748 | 0.2592 | 0.0547 | 0.0991 | 0.1054 |
| No log | 3.0 | 18 | 0.7718 | 0.2502 | 0.0447 | 0.0941 | 0.1114 |
| No log | 4.0 | 24 | 0.7677 | 0.2484 | 0.0448 | 0.0915 | 0.1121 |
| No log | 5.0 | 30 | 0.7692 | 0.2484 | 0.0450 | 0.0910 | 0.1124 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.3.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1