|
--- |
|
license: mit |
|
base_model: pyannote/segmentation-3.0 |
|
tags: |
|
- speaker-diarization |
|
- speaker-segmentation |
|
- generated_from_trainer |
|
datasets: |
|
- diarizers-community/callhome |
|
model-index: |
|
- name: speaker-segmentation-fine-tuned-callhome-eng |
|
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. --> |
|
|
|
# speaker-segmentation-fine-tuned-callhome-eng |
|
|
|
This model is a fine-tuned version of [pyannote/segmentation-3.0](https://huggingface.co/pyannote/segmentation-3.0) on the diarizers-community/callhome eng dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.4587 |
|
- Der: 0.1824 |
|
- False Alarm: 0.0587 |
|
- Missed Detection: 0.0707 |
|
- Confusion: 0.0529 |
|
|
|
## Model description |
|
This segmentation model has been trained on English data (Callhome) using [diarizers](https://github.com/huggingface/diarizers/tree/main). |
|
It can be loaded with two lines of code: |
|
|
|
```python |
|
from diarizers import SegmentationModel |
|
|
|
segmentation_model = SegmentationModel().from_pretrained('evie-8/speaker-segmentation-fine-tuned-callhome-eng') |
|
``` |
|
|
|
To use it within a pyannote speaker diarization pipeline, load the [pyannote/speaker-diarization-3.1](https://huggingface.co/pyannote/speaker-diarization-3.1) pipeline, and convert the model to a pyannote compatible format: |
|
|
|
```python |
|
|
|
from pyannote.audio import Pipeline |
|
import torch |
|
|
|
device = torch.device("cuda:0") if torch.cuda.is_available() else torch.device("cpu") |
|
|
|
|
|
# load the pre-trained pyannote pipeline |
|
pipeline = Pipeline.from_pretrained("pyannote/speaker-diarization-3.1") |
|
pipeline.to(device) |
|
|
|
# replace the segmentation model with your fine-tuned one |
|
model = segmentation_model.to_pyannote_model() |
|
pipeline._segmentation.model = model.to(device) |
|
``` |
|
|
|
You can now use the pipeline on audio examples: |
|
|
|
```python |
|
# load dataset example |
|
dataset = load_dataset("diarizers-community/callhome", "eng", split="data") |
|
sample = dataset[0]["audio"] |
|
|
|
# pre-process inputs |
|
sample["waveform"] = torch.from_numpy(sample.pop("array")[None, :]).to(device, dtype=model.dtype) |
|
sample["sample_rate"] = sample.pop("sampling_rate") |
|
|
|
# perform inference |
|
diarization = pipeline(sample) |
|
|
|
# dump the diarization output to disk using RTTM format |
|
with open("audio.rttm", "w") as rttm: |
|
diarization.write_rttm(rttm) |
|
``` |
|
|
|
|
|
## 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 | |
|
|:-------------:|:-----:|:----:|:---------------:|:------:|:-----------:|:----------------:|:---------:| |
|
| 0.4181 | 1.0 | 362 | 0.4878 | 0.1940 | 0.0577 | 0.0756 | 0.0607 | |
|
| 0.3931 | 2.0 | 724 | 0.4616 | 0.1827 | 0.0590 | 0.0718 | 0.0520 | |
|
| 0.3766 | 3.0 | 1086 | 0.4643 | 0.1826 | 0.0576 | 0.0723 | 0.0527 | |
|
| 0.3661 | 4.0 | 1448 | 0.4603 | 0.1832 | 0.0620 | 0.0682 | 0.0530 | |
|
| 0.3568 | 5.0 | 1810 | 0.4587 | 0.1824 | 0.0587 | 0.0707 | 0.0529 | |
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.41.2 |
|
- Pytorch 2.3.0+cu121 |
|
- Datasets 2.20.0 |
|
- Tokenizers 0.19.1 |
|
|
