mirror of
https://github.com/pyannote/pyannote-audio.git
synced 2026-11-20 13:07:51 +09:00
feat: preparing for pyannote.audio 3.0.0 (#1470)
This commit is contained in:
1 parent
76d86fa97a
commit
9a5a90251d
4 files changed
+66
-77
No files matched your search
+18
-25
@@ -1,24 +1,30 @@
|
||||
# Changelog
|
||||
|
||||
## Version 3.0 (xxxx-xx-xx)
|
||||
## Version 3.0.0 (2023-09-26)
|
||||
|
||||
### Highlights
|
||||
### Features and improvements
|
||||
|
||||
- *"Harder"*. Fixed [major reproducibility issue](https://github.com/pyannote/pyannote-audio/issues/1370) with Ampere (A100) NVIDIA GPUs
|
||||
In case you tried `pyannote.audio` pretrained pipelines in the past on Ampera (A100) NVIDIA GPUs
|
||||
and were disappointed by the accuracy, please give it another try with this new version.
|
||||
- "Better".
|
||||
- "Faster".
|
||||
- "Stronger".
|
||||
- feat(pipeline): send pipeline to device with `pipeline.to(device)`
|
||||
- feat(pipeline): add `return_embeddings` option to `SpeakerDiarization` pipeline
|
||||
- feat(pipeline): make `segmentation_batch_size` and `embedding_batch_size` mutable in `SpeakerDiarization` pipeline (they now default to `1`)
|
||||
- feat(pipeline): add progress hook to pipelines
|
||||
- feat(task): add [powerset](https://www.isca-speech.org/archive/interspeech_2023/plaquet23_interspeech.html) support to `SpeakerDiarization` task
|
||||
- feat(task): add support for multi-task models
|
||||
- feat(task): add support for label scope in speaker diarization task
|
||||
- feat(task): add support for missing classes in multi-label segmentation task
|
||||
- feat(model): add segmentation model based on torchaudio self-supervised representation
|
||||
- feat(pipeline): check version compatibility at load time
|
||||
- improve(task): load metadata as tensors rather than pyannote.core instances
|
||||
- improve(task): improve error message on missing specifications
|
||||
|
||||
### Breaking changes
|
||||
|
||||
- BREAKING(task): rename `Segmentation` task to `SpeakerDiarization`
|
||||
- BREAKING(task): remove support for variable chunk duration for segmentation tasks
|
||||
- BREAKING(pipeline): pipeline defaults to CPU (use `pipeline.to(device)`)
|
||||
- BREAKING(pipeline): remove `SpeakerSegmentation` pipeline (use `SpeakerDiarization` pipeline)
|
||||
- BREAKING(pipeline): remove support for `FINCHClustering` and `HiddenMarkovModelClustering`
|
||||
- BREAKING(pipeline): remove `segmentation_duration` parameter from `SpeakerDiarization` pipeline (defaults to `duration` of segmentation model)
|
||||
- BREAKING(task): remove support for variable chunk duration for segmentation tasks
|
||||
- BREAKING(pipeline): remove support for `FINCHClustering` and `HiddenMarkovModelClustering`
|
||||
- BREAKING(setup): drop support for Python 3.7
|
||||
- BREAKING(io): channels are now 0-indexed (used to be 1-indexed)
|
||||
- BREAKING(io): multi-channel audio is no longer downmixed to mono by default.
|
||||
@@ -29,21 +35,8 @@
|
||||
- BREAKING(model): get rid of (flaky) `Model.introspection`
|
||||
If, for some weird reason, you wrote some custom code based on that,
|
||||
you should instead rely on `Model.example_output`.
|
||||
- BREAKING(interactive): remove support for Prodigy recipes
|
||||
|
||||
### Features and improvements
|
||||
|
||||
- feat(task): add [powerset](https://www.isca-speech.org/archive/interspeech_2023/plaquet23_interspeech.html) support to `SpeakerDiarization` task
|
||||
- feat(task): add support for multi-task models
|
||||
- feat(task): add support for label scope in speaker diarization task
|
||||
- feat(task): add support for missing classes in multi-label segmentation task
|
||||
- feat(model): add segmentation model based on torchaudio self-supervised representation
|
||||
- feat(pipeline): send pipeline to device with `pipeline.to(device)`
|
||||
- feat(pipeline): add `return_embeddings` option to `SpeakerDiarization` pipeline
|
||||
- feat(pipeline): make `segmentation_batch_size` and `embedding_batch_size` mutable in `SpeakerDiarization` pipeline (they now default to `1`)
|
||||
- feat(pipeline): add progress hook to pipelines
|
||||
- feat(pipeline): check version compatibility at load time
|
||||
- improve(task): load metadata as tensors rather than pyannote.core instances
|
||||
- improve(task): improve error message on missing specifications
|
||||
|
||||
### Fixes and improvements
|
||||
|
||||
@@ -54,7 +47,7 @@
|
||||
- fix(task): fix support for "balance" option
|
||||
- improve(task): shorten and improve structure of Tensorboard tags
|
||||
|
||||
### Dependencies
|
||||
### Dependencies update
|
||||
|
||||
- setup: switch to torch 2.0+, torchaudio 2.0+, soundfile 0.12+, lightning 2.0+, torchmetrics 0.11+
|
||||
- setup: switch to pyannote.core 5.0+, pyannote.database 5.0+, and pyannote.pipeline 3.0+
|
||||
|
||||
@@ -1,30 +1,37 @@
|
||||
> [!IMPORTANT]
|
||||
> I propose (paid) scientific [consulting services](https://herve.niderb.fr/consulting.html) to companies willing to make the most of their data and open-source speech processing toolkits (and `pyannote` in particular).
|
||||
Using `pyannote.audio` open-source toolkit in production?
|
||||
Make the most of it thanks to our [consulting services](https://herve.niderb.fr/consulting.html).
|
||||
|
||||
# Speaker diarization with `pyannote.audio`
|
||||
# `pyannote.audio` speaker diarization toolkit
|
||||
|
||||
`pyannote.audio` is an open-source toolkit written in Python for speaker diarization. Based on [PyTorch](pytorch.org) machine learning framework, it provides a set of trainable end-to-end neural building blocks that can be combined and jointly optimized to build speaker diarization pipelines.
|
||||
`pyannote.audio` is an open-source toolkit written in Python for speaker diarization. Based on [PyTorch](pytorch.org) machine learning framework, it comes with state-of-the-art [pretrained models and pipelines](https://hf.co/pyannote), that can be further finetuned to your own data for even better performance.
|
||||
|
||||
<p align="center">
|
||||
<a href="https://www.youtube.com/watch?v=37R_R82lfwA"><img src="https://img.youtube.com/vi/37R_R82lfwA/0.jpg"></a>
|
||||
</p>
|
||||
|
||||
|
||||
## TL;DR [](https://colab.research.google.com/github/pyannote/pyannote-audio/blob/develop/tutorials/intro.ipynb)
|
||||
## TL;DR
|
||||
|
||||
1. Install [`pyannote.audio`](https://github.com/pyannote/pyannote-audio) `3.0` with `pip install pyannote.audio`
|
||||
2. Accept [`pyannote/segmentation-3.0`](https://hf.co/pyannote/segmentation-3.0) user conditions
|
||||
3. Accept [`pyannote/speaker-diarization-3.0`](https://hf.co/pyannote-speaker-diarization-3.0) user conditions
|
||||
4. Create access token at [`hf.co/settings/tokens`](https://hf.co/settings/tokens).
|
||||
|
||||
|
||||
```python
|
||||
# 1. visit hf.co/pyannote/speaker-diarization and hf.co/pyannote/segmentation and accept user conditions (only if requested)
|
||||
# 2. visit hf.co/settings/tokens to create an access token (only if you had to go through 1.)
|
||||
# 3. instantiate pretrained speaker diarization pipeline
|
||||
from pyannote.audio import Pipeline
|
||||
pipeline = Pipeline.from_pretrained("pyannote/speaker-diarization",
|
||||
use_auth_token="ACCESS_TOKEN_GOES_HERE")
|
||||
pipeline = Pipeline.from_pretrained(
|
||||
"pyannote/speaker-diarization-3.0",
|
||||
use_auth_token="HUGGINGFACE_ACCESS_TOKEN_GOES_HERE")
|
||||
|
||||
# 4. apply pretrained pipeline
|
||||
# send pipeline to GPU (when available)
|
||||
import torch
|
||||
pipeline.to(torch.device("cuda"))
|
||||
|
||||
# apply pretrained pipeline
|
||||
diarization = pipeline("audio.wav")
|
||||
|
||||
# 5. print the result
|
||||
# print the result
|
||||
for turn, _, speaker in diarization.itertracks(yield_label=True):
|
||||
print(f"start={turn.start:.1f}s stop={turn.end:.1f}s speaker_{speaker}")
|
||||
# start=0.2s stop=1.5s speaker_0
|
||||
@@ -39,16 +46,7 @@ for turn, _, speaker in diarization.itertracks(yield_label=True):
|
||||
- :exploding_head: state-of-the-art performance (see [Benchmark](#benchmark))
|
||||
- :snake: Python-first API
|
||||
- :zap: multi-GPU training with [pytorch-lightning](https://pytorchlightning.ai/)
|
||||
- :control_knobs: data augmentation with [torch-audiomentations](https://github.com/asteroid-team/torch-audiomentations)
|
||||
|
||||
## Installation
|
||||
|
||||
Only Python 3.8+ is supported.
|
||||
|
||||
```bash
|
||||
# install from develop branch
|
||||
pip install -qq https://github.com/pyannote/pyannote-audio/archive/refs/heads/develop.zip
|
||||
```
|
||||
|
||||
## Documentation
|
||||
|
||||
@@ -72,53 +70,50 @@ pip install -qq https://github.com/pyannote/pyannote-audio/archive/refs/heads/de
|
||||
- 2022-12-02 > ["How I reached 1st place at Ego4D 2022, 1st place at Albayzin 2022, and 6th place at VoxSRC 2022 speaker diarization challenges"](tutorials/adapting_pretrained_pipeline.ipynb)
|
||||
- 2022-10-23 > ["One speaker segmentation model to rule them all"](https://herve.niderb.fr/fastpages/2022/10/23/One-speaker-segmentation-model-to-rule-them-all)
|
||||
- 2021-08-05 > ["Streaming voice activity detection with pyannote.audio"](https://herve.niderb.fr/fastpages/2021/08/05/Streaming-voice-activity-detection-with-pyannote.html)
|
||||
- Miscellaneous
|
||||
- [Training with `pyannote-audio-train` command line tool](tutorials/training_with_cli.md)
|
||||
- [Speaker verification](tutorials/speaker_verification.ipynb)
|
||||
- Visualization and debugging
|
||||
- Videos
|
||||
- [Introduction to speaker diarization](https://umotion.univ-lemans.fr/video/9513-speech-segmentation-and-speaker-diarization/) / JSALT 2023 summer school / 90 min
|
||||
- [Speaker segmentation model](https://www.youtube.com/watch?v=wDH2rvkjymY) / Interspeech 2021 / 3 min
|
||||
- [First releaase of pyannote.audio](https://www.youtube.com/watch?v=37R_R82lfwA) / ICASSP 2020 / 8 min
|
||||
|
||||
## Benchmark
|
||||
|
||||
Out of the box, `pyannote.audio` default speaker diarization [pipeline](https://hf.co/pyannote/speaker-diarization) is expected to be much better (and faster) in v2.x than in v1.1. Those numbers are diarization error rates (in %)
|
||||
Out of the box, `pyannote.audio` speaker diarization [pipeline](https://hf.co/pyannote/speaker-diarization-3.0) v3.0 is expected to be much better (and faster) than v2.x.
|
||||
Those numbers are diarization error rates (in %):
|
||||
|
||||
| Dataset \ Version | v1.1 | v2.0 | v2.1.1 (finetuned) |
|
||||
| ---------------------- | ---- | ---- | ------------------ |
|
||||
| AISHELL-4 | - | 14.6 | 14.1 (14.5) |
|
||||
| AliMeeting (channel 1) | - | - | 27.4 (23.8) |
|
||||
| AMI (IHM) | 29.7 | 18.2 | 18.9 (18.5) |
|
||||
| AMI (SDM) | - | 29.0 | 27.1 (22.2) |
|
||||
| CALLHOME (part2) | - | 30.2 | 32.4 (29.3) |
|
||||
| DIHARD 3 (full) | 29.2 | 21.0 | 26.9 (21.9) |
|
||||
| VoxConverse (v0.3) | 21.5 | 12.6 | 11.2 (10.7) |
|
||||
| REPERE (phase2) | - | 12.6 | 8.2 ( 8.3) |
|
||||
| This American Life | - | - | 20.8 (15.2) |
|
||||
| Dataset \ Version | v1.1 | v2.0 | [v2.1](https://hf.co/pyannote/speaker-diarization-2.1) | [v3.0](https://hf.co/pyannote/speaker-diarization-3.0) | <a href="mailto:herve-at-niderb-dot-fr?subject=Premium pyannote.audio pipeline&body=Looks like I got your attention! Drop me an email for more details. Hervé.">Premium</a> |
|
||||
| ---------------------- | ---- | ---- | ------ | ------ | --------- |
|
||||
| AISHELL-4 | - | 14.6 | 14.1 | 12.3 | 12.3 |
|
||||
| AliMeeting (channel 1) | - | - | 27.4 | 24.3 | 19.4 |
|
||||
| AMI (IHM) | 29.7 | 18.2 | 18.9 | 19.0 | 16.7 |
|
||||
| AMI (SDM) | - | 29.0 | 27.1 | 22.2 | 20.1 |
|
||||
| AVA-AVD | - | - | - | 49.1 | 42.7 |
|
||||
| DIHARD 3 (full) | 29.2 | 21.0 | 26.9 | 21.7 | 17.0 |
|
||||
| MSDWild | - | - | - | 24.6 | 20.4 |
|
||||
| REPERE (phase2) | - | 12.6 | 8.2 | 7.8 | 7.8 |
|
||||
| VoxConverse (v0.3) | 21.5 | 12.6 | 11.2 | 11.3 | 9.5 |
|
||||
|
||||
## Citations
|
||||
|
||||
If you use `pyannote.audio` please use the following citations:
|
||||
|
||||
```bibtex
|
||||
@inproceedings{Bredin2020,
|
||||
Title = {{pyannote.audio: neural building blocks for speaker diarization}},
|
||||
Author = {{Bredin}, Herv{\'e} and {Yin}, Ruiqing and {Coria}, Juan Manuel and {Gelly}, Gregory and {Korshunov}, Pavel and {Lavechin}, Marvin and {Fustes}, Diego and {Titeux}, Hadrien and {Bouaziz}, Wassim and {Gill}, Marie-Philippe},
|
||||
Booktitle = {ICASSP 2020, IEEE International Conference on Acoustics, Speech, and Signal Processing},
|
||||
Year = {2020},
|
||||
@inproceedings{Plaquet23,
|
||||
author={Alexis Plaquet and Hervé Bredin},
|
||||
title={{Powerset multi-class cross entropy loss for neural speaker diarization}},
|
||||
year=2023,
|
||||
booktitle={Proc. INTERSPEECH 2023},
|
||||
}
|
||||
```
|
||||
|
||||
```bibtex
|
||||
@inproceedings{Bredin2021,
|
||||
Title = {{End-to-end speaker segmentation for overlap-aware resegmentation}},
|
||||
Author = {{Bredin}, Herv{\'e} and {Laurent}, Antoine},
|
||||
Booktitle = {Proc. Interspeech 2021},
|
||||
Year = {2021},
|
||||
@inproceedings{Bredin23,
|
||||
author={Hervé Bredin},
|
||||
title={{pyannote.audio 2.1 speaker diarization pipeline: principle, benchmark, and recipe}},
|
||||
year=2023,
|
||||
booktitle={Proc. INTERSPEECH 2023},
|
||||
}
|
||||
```
|
||||
|
||||
## Support
|
||||
|
||||
For commercial enquiries and scientific consulting, please contact [me](mailto:herve@niderb.fr).
|
||||
|
||||
## Development
|
||||
|
||||
The commands below will setup pre-commit hooks and packages needed for developing the `pyannote.audio` library.
|
||||
|
||||
@@ -3,6 +3,7 @@ einops >=0.6.0
|
||||
huggingface_hub >= 0.13.0
|
||||
lightning >= 2.0.1
|
||||
omegaconf >=2.1,<3.0
|
||||
onnxruntime >= 1.16.0
|
||||
pyannote.core >= 5.0.0
|
||||
pyannote.database >= 5.0.1
|
||||
pyannote.metrics >= 3.2
|
||||
|
||||
+1
-1
@@ -1 +1 @@
|
||||
2.1.1
|
||||
3.0.0
|
||||
Reference in new issue
Block a user