xuhaorran/WSYue-ASR-eval

WSYue-ASR-eval是一个专门为评估粤语自动语音识别(ASR)系统设计的基准数据集。该数据集针对粤语在语音识别中的独特语言特征，旨在评估模型在不同长度、领域和语言现象下的表现。测试集标注由北京AISHELL科技有限公司提供，具有多轮人工标注、包含文本转录、情感、年龄、性别等丰富标签，涵盖粤英代码转换和多领域条件，支持不同语音长度的全面评估。数据集包含短语音(0-10秒)和长语音(10-30秒)两个子集，总计11.4小时，涉及2861名短语音说话人和838名长语音说话人，覆盖多样化的说话人和场景。

WSYue-ASR-eval is a benchmark specifically designed for evaluating Cantonese ASR systems. It is tailored to assess model performance across diverse lengths, domains, and linguistic phenomena of Cantonese speech. The test set annotations are provided by Beijing AISHELL Technology Co., Ltd., featuring multiple rounds of manual labeling, rich tags including text transcription, emotion, age, and gender, covering Cantonese-English code-switching and multi-domain conditions, enabling comprehensive evaluation across varying speech lengths. The dataset contains two subsets: Short (0-10s) and Long (10-30s), totaling 11.4 hours with 2861 short-utterance speakers and 838 long-utterance speakers, covering diverse speakers and scenarios.