|
--- |
|
library_name: transformers |
|
language: |
|
- spa |
|
tags: |
|
- speaker-diarization |
|
- speaker-segmentation |
|
- generated_from_trainer |
|
datasets: |
|
- pyannote/segmentation |
|
model-index: |
|
- name: segmentation-3.0 |
|
results: [] |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# segmentation-3.0 |
|
|
|
This model is a fine-tuned version of [](https://huggingface.co/) on the pyannote/segmentation dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.6638 |
|
- Der: 0.2932 |
|
- False Alarm: 0.2540 |
|
- Missed Detection: 0.0387 |
|
- Confusion: 0.0004 |
|
|
|
## 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: 64 |
|
- eval_batch_size: 64 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- num_epochs: 5 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Der | False Alarm | Missed Detection | Confusion | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:|:-----------:|:----------------:|:---------:| |
|
| 0.3145 | 1.0 | 282 | 0.5751 | 0.2944 | 0.2487 | 0.0454 | 0.0003 | |
|
| 0.3087 | 2.0 | 564 | 0.5957 | 0.2912 | 0.2462 | 0.0440 | 0.0010 | |
|
| 0.2905 | 3.0 | 846 | 0.6614 | 0.2970 | 0.2627 | 0.0333 | 0.0010 | |
|
| 0.2733 | 4.0 | 1128 | 0.6626 | 0.2940 | 0.2558 | 0.0378 | 0.0004 | |
|
| 0.2672 | 5.0 | 1410 | 0.6638 | 0.2932 | 0.2540 | 0.0387 | 0.0004 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.45.1 |
|
- Pytorch 2.4.1 |
|
- Datasets 3.0.1 |
|
- Tokenizers 0.20.0 |
|
|
