speaker-segmentation-fine-tuned-ita
This model is a fine-tuned version of pyannote/segmentation-3.0 on the Obiactum/synthetic-speaker-diarization-dataset-ita dataset. It achieves the following results on the evaluation set:
- Loss: 0.3340
- Model Preparation Time: 0.006
- Der: 0.1101
- False Alarm: 0.0132
- Missed Detection: 0.0192
- Confusion: 0.0777
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 | Model Preparation Time | Der | False Alarm | Missed Detection | Confusion |
---|---|---|---|---|---|---|---|---|
0.3773 | 1.0 | 472 | 0.3665 | 0.006 | 0.1227 | 0.0151 | 0.0213 | 0.0863 |
0.3264 | 2.0 | 944 | 0.3458 | 0.006 | 0.1172 | 0.0135 | 0.0202 | 0.0835 |
0.3074 | 3.0 | 1416 | 0.3463 | 0.006 | 0.1162 | 0.0144 | 0.0188 | 0.0830 |
0.2837 | 4.0 | 1888 | 0.3338 | 0.006 | 0.1110 | 0.0136 | 0.0188 | 0.0786 |
0.2852 | 5.0 | 2360 | 0.3340 | 0.006 | 0.1101 | 0.0132 | 0.0192 | 0.0777 |
Framework versions
- Transformers 4.43.3
- Pytorch 2.4.0+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1
- Downloads last month
- 2
Model tree for Obiactum/speaker-segmentation-fine-tuned-ita
Base model
pyannote/segmentation-3.0