metadata
library_name: transformers
license: mit
base_model: pyannote/segmentation-3.0
tags:
- speaker-diarization
- speaker-segmentation
- generated_from_trainer
datasets:
- Venkatesh4342/pyannote-hindi-diarization
model-index:
- name: speaker-segmentation-fine-tuned-hi
results: []
speaker-segmentation-fine-tuned-hi
This model is a fine-tuned version of pyannote/segmentation-3.0 on the Venkatesh4342/pyannote-hindi-diarization dataset. It achieves the following results on the evaluation set:
- Loss: 0.0767
- Model Preparation Time: 0.0076
- Der: 0.0269
- False Alarm: 0.0116
- Missed Detection: 0.0113
- Confusion: 0.0040
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 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: 5.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Model Preparation Time | Der | False Alarm | Missed Detection | Confusion |
|---|---|---|---|---|---|---|---|---|
| 0.1348 | 1.0 | 2674 | 0.1155 | 0.0076 | 0.0404 | 0.0167 | 0.0145 | 0.0093 |
| 0.0936 | 2.0 | 5348 | 0.0908 | 0.0076 | 0.0319 | 0.0111 | 0.0155 | 0.0053 |
| 0.0838 | 3.0 | 8022 | 0.0803 | 0.0076 | 0.0283 | 0.0111 | 0.0126 | 0.0045 |
| 0.0751 | 4.0 | 10696 | 0.0779 | 0.0076 | 0.0275 | 0.0122 | 0.0111 | 0.0042 |
| 0.0871 | 5.0 | 13370 | 0.0767 | 0.0076 | 0.0269 | 0.0116 | 0.0113 | 0.0040 |
Framework versions
- Transformers 4.49.0
- Pytorch 2.6.0+cu124
- Datasets 3.4.1
- Tokenizers 0.21.1