401小时川渝方言自然对话（双语标注）手机采集语音数据

川渝话自然对话（双语标注）手机采集语音数据，基于30余个常见主题来模拟录制。此数据集标注了方言及普通话两版文本内容、句时间戳、说话人身份、性别等多种属性，由来自川渝地区的方言使用者录制，准确性高，为语音识别相关研究及应用提供了丰富的资源，经多家AI公司验证：有助于模型面对真实世界的多样性时能够表现出色。我们严格遵循数据保护法规和隐私规定，确保数据采集、存储和使用的过程中维护用户的隐私和合法权益，所有数据均遵循GDPR, CCPA, PIPL。Natural Conversations in Sichuan-Chongqing Dialect (Bilingual Annotated): Mobile-collected Speech Dataset. This dataset was recorded based on over 30 common conversational themes, and is annotated with multiple attributes including bilingual transcriptions (Sichuan-Chongqing dialect and Standard Mandarin versions), sentence-level timestamps, speaker identity, gender, and more. Recorded by native dialect speakers from Sichuan and Chongqing regions, the dataset boasts high accuracy, serving as a rich resource for speech recognition-related research and applications. Verified by multiple AI enterprises, it enables models to perform well when facing the diversity of real-world scenarios. We strictly abide by data protection laws and privacy regulations to safeguard users' privacy and legitimate rights and interests during data collection, storage and utilization. All datasets comply with GDPR, CCPA and PIPL.

搜集汇总

背景与挑战

背景概述

该数据集包含401小时川渝方言自然对话语音，由486名来自川渝地区的说话人通过手机录制，覆盖多个年龄段和均衡性别比例。数据标注了方言及普通话双版本文本、句时间戳和说话人身份等信息，基于30余个常见主题模拟真实对话，准确率高，适用于语音识别和声纹识别研究与应用。