数据堂—155人马来语手机采集数据_引导

马来语手机采集数据_引导是由155位马来西亚本地人员参与采集。录音环境安静，录音内容包括车载、家居、语音助手等多个领域。每人约50句，有效数据时长7小时，所有文本由人工转写，准确率高

The Malay-language mobile-collected dataset with guided prompts was recruited and collected by 155 local Malaysian participants. The audio recordings were conducted in quiet environments, covering multiple domains including in-car scenarios, home environments, voice assistant interactions and others. Each participant contributed approximately 50 utterances, with a total valid audio duration of 7 hours. All transcripts were manually produced with high accuracy.

搜集汇总

背景与挑战

背景概述

该数据集专为马来语语音识别模型测试设计，由155名马来西亚参与者通过手机采集，总有效时长为7小时，覆盖车内、家庭和语音助手等多个安静环境下的场景，所有文本均经过人工转录确保准确性。