WenetSpeech

Visit Tool

WenetSpeech is a Data & Analytics tool that provides a 10000+ hour dataset for Chinese speech recognition. It is a multi-domain corpus designed for training and evaluating speech recognition models.

Claim this tool

No Views Yet

At a glance

Pricing

Open Source

Free tier

Yes

API

Skill level

Technical

About

What is WenetSpeech?

WenetSpeech offers a comprehensive 10000+ hour multi-domain Chinese corpus specifically designed for speech recognition tasks. This extensive dataset is compiled from YouTube and Podcast sources, utilizing both Optical Character Recognition (OCR) and Automatic Speech Recognition (ASR) techniques for labeling. To ensure high quality, the corpus undergoes a novel end-to-end label error detection method for validation and filtering. It categorizes data into High Label, Weak Label, and Unlabel sets, suitable for supervised, semi-supervised, or unsupervised training. The dataset also provides various training subsets (S, M, L) and evaluation sets (DEV, TEST_NET, TEST_MEETING) to support diverse ASR system development and benchmarking. Access to the dataset requires visiting the official website, agreeing to the license, and obtaining a password.

Best used for

Ideal for researchers and developers who need to train robust Chinese Automatic Speech Recognition (ASR) models, evaluate existing ASR systems, and conduct experiments with large-scale speech data. Especially valuable for those requiring high-quality, multi-domain Chinese speech data for supervised, semi-supervised, or unsupervised learning.

Common actions

train speech models

evaluate ASR systems

access Chinese speech data

benchmark speech recognition

"AI Agents"face swappinggithub copilotworkflowsdeepfakeautomated workflowlow-code/no-codeopen-sourcecollaboration

Capabilities

Key features

10000+ hours Chinese speech
Multi-domain corpus
OCR/ASR labeling
Label error detection
Categorized data sets
Multiple training subsets
Dedicated evaluation sets

Target Audience

ai researchersasr developersdata scientistsmachine learning engineers

Integrations

Not yet documented

Pricing & Plans

Open Source

Free

FAQs

How can I access the WenetSpeech dataset?

To access the WenetSpeech dataset, you need to visit the official website, read and agree to the license terms, and then apply for a password. Once you have the password, you can use the provided bash scripts to download the data.

What types of data are included in WenetSpeech?

WenetSpeech includes over 10000 hours of Chinese speech data categorized into High Label, Weak Label, and Unlabel sets. The data is sourced from YouTube and Podcasts, covering various domains like drama, news, interviews, and more, suitable for different training paradigms.

Are there different training subsets available?

Yes, WenetSpeech provides three training subsets: S (100 hours), M (1000 hours), and L (10005 hours), all with 1.0 confidence or higher. These subsets allow users to build ASR systems on different data scales, catering to various research and development needs.

Trending

Subcategories trending in Data & Analytics

Business Intelligence Predictive Analytics Real-Time Analytics Market Research Data Cleaning & Prep Data Pipelines & Integration

Trending

Also listed in

This tool also appears in

Research & Education › Academic Research Content & Design › Audio & Music Coding & Development › Open Source & Models

Explore

Browse AI tools by category

Content & Design Productivity & Business Coding & Development AI Agents & Automation Research & Education Wellness & Lifestyle Career Development Marketing & Growth Data & Analytics Customer Support & CX Finance E-commerce