📢: Good news! 21,800 hours of multi-label Cantonese speech data are also available at  WenetSpeech-Yue                                                                                                              📢: Good news! 21,800 hours of multi-label Cantonese speech data are also available at  WenetSpeech-Yue                                                                                                              📢: Good news! 21,800 hours of multi-label Cantonese speech data are also available at  WenetSpeech-Yue

WenetSpeech-Chuan: A Large-Scale Sichuanese Corpus With Rich Annotation For Dialectal Speech Processing

📑 Paper    |    🐙 GitHub    |    🤗 HuggingFace
🎤 Demo Page    |    💬 Contact Us

Abstract

The scarcity of large-scale, open-source data for dialects severely hinders progress in speech technology, a challenge particularly acute for the widely spoken Sichuanese dialects of Chinese. To address this critical gap, we introduce WenetSpeech-Chuan, a 10,000-hour, richly annotated corpus constructed using our novel Chuan-Pipeline, a complete data processing framework for dialectal speech. To facilitate rigorous evaluation and demonstrate the corpus's effectiveness, we also release high-quality ASR and TTS benchmarks, WenetSpeech-Chuan-Eval, with manually verified transcriptions. Experiments show that models trained on WenetSpeech-Chuan achieve state-of-the-art performance among open-source systems and demonstrate results comparable to commercial services. As the largest open-source corpus for Sichuanese dialects, WenetSpeech-Chuan not only lowers the barrier to research in dialectal speech processing but also plays a crucial role in promoting AI equity and mitigating bias in speech technologies. The corpus, benchmarks, models, and receipts are publicly available. Demos can be found in the supplementary material.

Promotional video

Chuan-Pipeline

Chuan-Pipeline is an automated pipeline specifically designed for building large-scale Cantonese datasets with multi-dimensional annotation. It consists of six components: (A) Audio Collection, (B) Speaker Attributes Annotation, (C) Speech Quality Annotation, (D) Automatic Speech Recognition, (E) Text Post-Processing, and (F) Recognizer Output Voting. The figure below provides an overview of the Chuan-Pipeline.

WenetSpeech-Chuan

Dataset Overview

Data Samples

ASR Leaderboard

Leaderboard shows ASR Results (CER%↓) on Sichuanese Datasets.

Note: Bold indicates best performance, underlined indicates second-best performance, and light green background indicates models finetuned on a high-quality internal corpus (to show the system's potential as a foundation model).

TTS Evaluation

TTS Demo

Model Comparison

CosyVoice2-WSC is a finetuned CosyVoice2 model using WenetSpeech-Chuan.

Llasa-1B-WSC is a finetuned Llasa-1B model using WenetSpeech-Chuan.

CosyVoice2-WSC-SFT

Further supervised fine-tune CosyVoice2-WSC with 100 hours of internal high-quality data from two fixed speakers.