metadata
tags:
- pyannote
- pyannote-audio
- pyannote-audio-model
- audio
- voice
- speech
- speaker
- speaker-recognition
- speaker-verification
- speaker-identification
- speaker-embedding
datasets:
- voxceleb
license: mit
inference: false
extra_gated_prompt: >-
The collected information will help acquire a better knowledge of
pyannote.audio userbase and help its maintainers apply for grants to improve
it further. If you are an academic researcher, please cite the relevant papers
in your own publications using the model. If you work for a company, please
consider contributing back to pyannote.audio development (e.g. through
unrestricted gifts). We also provide scientific consulting services around
speaker diarization and machine listening.
extra_gated_fields:
Company/university: text
Website: text
I plan to use this model for (task, type of audio data, etc): text
Using this open-source model in production?
Make the most of it thanks to our consulting services.
🎹 Speaker embedding
Relies on pyannote.audio 2.1: see installation instructions.
This model is based on the canonical x-vector TDNN-based architecture, but with filter banks replaced with trainable SincNet features. See XVectorSincNet architecture for implementation details.
Basic usage
# 1. visit hf.co/pyannote/embedding and accept user conditions
# 2. visit hf.co/settings/tokens to create an access token
# 3. instantiate pretrained model
from pyannote.audio import Model
model = Model.from_pretrained("pyannote/embedding",
use_auth_token="ACCESS_TOKEN_GOES_HERE")
from pyannote.audio import Inference
inference = Inference(model, window="whole")
embedding1 = inference("speaker1.wav")
embedding2 = inference("speaker2.wav")
# `embeddingX` is (1 x D) numpy array extracted from the file as a whole.
from scipy.spatial.distance import cdist
distance = cdist(embedding1, embedding2, metric="cosine")[0,0]
# `distance` is a `float` describing how dissimilar speakers 1 and 2 are.
Using cosine distance directly, this model reaches 2.8% equal error rate (EER) on VoxCeleb 1 test set.
This is without voice activity detection (VAD) nor probabilistic linear discriminant analysis (PLDA).
Expect even better results when adding one of those.
Advanced usage
Running on GPU
import torch
inference.to(torch.device("cuda"))
embedding = inference("audio.wav")
Extract embedding from an excerpt
from pyannote.audio import Inference
from pyannote.core import Segment
inference = Inference(model, window="whole")
excerpt = Segment(13.37, 19.81)
embedding = inference.crop("audio.wav", excerpt)
# `embedding` is (1 x D) numpy array extracted from the file excerpt.
Extract embeddings using a sliding window
from pyannote.audio import Inference
inference = Inference(model, window="sliding",
duration=3.0, step=1.0)
embeddings = inference("audio.wav")
# `embeddings` is a (N x D) pyannote.core.SlidingWindowFeature
# `embeddings[i]` is the embedding of the ith position of the
# sliding window, i.e. from [i * step, i * step + duration].
Citation
@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},
Address = {Barcelona, Spain},
Month = {May},
Year = {2020},
}
@inproceedings{Coria2020,
author="Coria, Juan M. and Bredin, Herv{\'e} and Ghannay, Sahar and Rosset, Sophie",
editor="Espinosa-Anke, Luis and Mart{\'i}n-Vide, Carlos and Spasi{\'{c}}, Irena",
title="{A Comparison of Metric Learning Loss Functions for End-To-End Speaker Verification}",
booktitle="Statistical Language and Speech Processing",
year="2020",
publisher="Springer International Publishing",
pages="137--148",
isbn="978-3-030-59430-5"
}