deep-privacy / sidekit

For further release go to: https://git-lium.univ-lemans.fr/speaker/sidekit

SIDEKIT

SIDEKIT is an open source package for Speaker and Language recognition. This repo provides the python/pytorch implementation the Deep Neural Network based Automatic Speaker Recognition systems and training losses of: Larcher/sidekit

Authors: Anthony Larcher & Kong Aik Lee & Sylvain Meignier

---

⚠️ For further release go to: https://git-lium.univ-lemans.fr/speaker/sidekit

Installation

git clone https://github.com/deep-privacy/sidekit
cd sidekit
# you might need to adjust $CUDAROOT in ./install.sh
# to match your cuda config. default /usr/local/cuda
./install.sh

Usage

For kaldi-like wav.scp

# activate the miniconda venv
. ./env.sh

# download trained model
# Model: HalfResNet34
# Trained with Loss: Large margin arc distance
# Trained on: VoxCeleb1 & 2
# Test EER on vox1-O: `1.20 %`
wget https://github.com/deep-privacy/sidekit/releases/download/sidekit_v0.1/best_halp_clr_adam_aam0.2_30_b256_vox12.pt_epoch71

# wav.scp to extract (kaldi-like)
cd egs/examples_decode

# extract and store the x-vectors in a scp,ark file
extract_xvectors.py --model ../../best_halp_clr_adam_aam0.2_30_b256_vox12.pt_epoch71 \
        --wav-scp ./wav_example.scp --out-scp ./x-vector.scp # the "--vad" flag can be used to remove non speech

For Python

import torch
import torchaudio
from sidekit.nnet.xvector import Xtractor

model_path = "./best_halp_clr_adam_aam0.2_30_b256_vox12.pt_epoch71"
device = "cuda" if torch.cuda.is_available() else "cpu"

model_config = torch.load(model_path, map_location=device)

model_opts = model_config["model_archi"]
if "embedding_size" not in model_opts:
    model_opts["embedding_size"] = 256
xtractor = Xtractor(model_config["speaker_number"],
                 model_archi=model_opts["model_type"],
                 loss=model_opts["loss"]["type"],
                 embedding_size=model_opts["embedding_size"])

xtractor.load_state_dict(model_config["model_state_dict"], strict=True)
xtractor = xtractor.to(device)
xtractor.eval()

wav_tensor, sample_rate = torchaudio.load("egs/examples_decode/1272-128104-0000.wav")
assert sample_rate == 16000 # same as in Sidekit training
# You have to apply VAD on your own! (Check extract_xvectors.py for an example)
_, vec = xtractor(wav_tensor.to(device), is_eval=True)
print(vec.shape)