An evaluation toolkit for voice conversion models.

Last update: Aug 29, 2022

Overview

Voice-conversion-evaluation

An evaluation toolkit for voice conversion models.

Sample test pair

Generate the metadata for evaluating models.
The directory of parsers contains several available corpus parsers.

  python sampler.py [name of source corpus] [path of source dir] [name of target corpus] [path of target dir] -n [number of samples] -nt [number of target utterances] -o [path of output dir]

The pairs of metadata are sorted by src_second for long to short.
The metadata contains:

source_corpus: The name of the source corpus.
source_corpus_speaker_number: The number of speaker in source corpus.
source_random_seed: Random seed used for sampling source utterance.
target_corpus: The name of the target corpus.
target_corpus_speaker_number: The number of speaker in target corpus.
target_random_seed: Random seed used for sampling target utterances.
n_samples: number of samples
n_target_samples: number of target utterances
pairs: List of evaluating pairs
- source_speaker: The name of the source speaker.
- target_speaker: The name of the target speaker.
- src_utt: The relative path of the source utterance, which is relative to the source dir.
- tgt_utts: List of the relative path of target utterances, which is relative to the target dir.
- content: The content of the source utterance.
- src_second: The second of the source utterance.
- converted: The entry does not appear when use sampler, you need to add the relative path for your converted output.

Metrics

The metrics include automatic mean opinion score assessment, character error rate, and speaker verification acceptance rate.

Automatic mean opinion score assessment
- Ensemble several MBNet which is implemented by sky1456723.
```
  python calculate_objective_metric.py -d [data_dir] -r metrics/mean_opinion_score
```
Character error rate:
- Use the automatic speech recognition model provided by Hugging Face.
- The word error rate on Librispeech test-other is 3.9.
```
  python calculate_objective_metric.py -d [data_dir] -r metrics/character_error_rate
```
Speaker verification acceptance rate:
- You can calculate the threshold by metrics/speaker_verification/equal_error_rate/.
- And some pre-calculated thresholds are in metrics/speaker_verification/equal_error_rate/threshold.yaml.
```
  python calculate_objective_metric.py -d [data_dir] -r metrics/speaker_verification -t [target_dir] -th [threshold path]
```

Voice assistant - Voice assistant with python

🌐 Python Voice Assistant 🌵 - User's greeting 🌵 - Writing tasks to todo-list ?

10 Dec 26, 2022

MMdnn is a set of tools to help users inter-operate among different deep learning frameworks. E.g. model conversion and visualization. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML.

MMdnn MMdnn is a comprehensive and cross-framework tool to convert, visualize and diagnose deep learning (DL) models. The "MM" stands for model manage

5.7k Jan 9, 2023

SIMULEVAL A General Evaluation Toolkit for Simultaneous Translation

SimulEval SimulEval is a general evaluation framework for simultaneous translation on text and speech. Requirement python = 3.7.0 Installation git cl

48 Dec 28, 2022

Microsoft Cognitive Toolkit (CNTK), an open source deep-learning toolkit

CNTK Chat Windows build status Linux build status The Microsoft Cognitive Toolkit (https://cntk.ai) is a unified deep learning toolkit that describes

17.3k Dec 29, 2022

Microsoft Cognitive Toolkit (CNTK), an open source deep-learning toolkit

CNTK Chat Windows build status Linux build status The Microsoft Cognitive Toolkit (https://cntk.ai) is a unified deep learning toolkit that describes

17k Feb 11, 2021

This project is a loose implementation of paper "Algorithmic Financial Trading with Deep Convolutional Neural Networks: Time Series to Image Conversion Approach"

Stock Market Buy/Sell/Hold prediction Using convolutional Neural Network This repo is an attempt to implement the research paper titled "Algorithmic F

136 Dec 28, 2022

An evaluation toolkit for voice conversion models.

Related tags

Overview

Voice-conversion-evaluation

Sample test pair

Metrics

You might also like...

Voice assistant - Voice assistant with python

MMdnn is a set of tools to help users inter-operate among different deep learning frameworks. E.g. model conversion and visualization. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML.

SIMULEVAL A General Evaluation Toolkit for Simultaneous Translation

Microsoft Cognitive Toolkit (CNTK), an open source deep-learning toolkit

Microsoft Cognitive Toolkit (CNTK), an open source deep-learning toolkit

This project is a loose implementation of paper "Algorithmic Financial Trading with Deep Convolutional Neural Networks: Time Series to Image Conversion Approach"

PlenOctrees: NeRF-SH Training & Conversion

Implementation for "Manga Filling Style Conversion with Screentone Variational Autoencoder" (SIGGRAPH ASIA 2020 issue)

[ICCV2021] IICNet: A Generic Framework for Reversible Image Conversion

Releases(checkpoints)

checkpoints(May 17, 2021)

Owner

This repository provides an efficient PyTorch-based library for training deep models.

The PyTorch implementation of paper REST: Debiased Social Recommendation via Reconstructing Exposure Strategies

A little Python application to auto tag your photos with the power of machine learning.

Unofficial PyTorch implementation of SimCLR by Google Brain

[NeurIPS 2021] Low-Rank Subspaces in GANs

Pretrained Pytorch face detection (MTCNN) and recognition (InceptionResnet) models

Generating synthetic mobility data for a realistic population with RNNs to improve utility and privacy

Neural Surface Maps

Code for the SIGGRAPH 2021 paper "Consistent Depth of Moving Objects in Video".

Official git for "CTAB-GAN: Effective Table Data Synthesizing"

本项目是一个带有前端界面的垃圾分类项目，加载了训练好的模型参数，模型为efficientnetb4，暂时为40分类问题。

ML-based medical imaging using Azure

Predicting Tweet Sentiment Maching Learning and streamlit

A tool to prepare websites grabbed with wget for local viewing.

This code reproduces the results of the paper, "Measuring Data Leakage in Machine-Learning Models with Fisher Information"

Single object tracking and segmentation.

Data and analysis code for an MS on SK VOC genomes phenotyping/neutralisation assays

Compute execution plan: A DAG representation of work that you want to get done. Individual nodes of the DAG could be simple python or shell tasks or complex deeply nested parallel branches or embedded DAGs themselves.

Official implementation of the Neurips 2021 paper Searching Parameterized AP Loss for Object Detection.

Large Scale Multi-Illuminant (LSMI) Dataset for Developing White Balance Algorithm under Mixed Illumination