普通话精标数据集

该识别数据在安静的办公室、住宅、录音棚和嘈杂的电视、汽车、娱乐、住宅、厨房、街道、办公室、地铁、餐馆、学校环境中完成录制。包含73361位发音人，10万小时数据。数据包含智能客服，智能导航等多种类型，可应用于客户服务领域，教育，医疗，汽车导航，医疗健康，游戏娱乐等多种场景的模型训练，提升客户使用体验This speech recognition dataset was recorded across diverse acoustic environments, including quiet offices, residences, and recording studios, as well as noisy scenarios such as television playback contexts, cars, entertainment venues, residences, kitchens, streets, offices, subway stations, restaurants, and schools. It includes audio data from 73,361 speakers, totaling 100,000 hours. The dataset covers various content types such as intelligent customer service and intelligent navigation scenarios, and can be applied to model training in multiple scenarios including customer service, education, healthcare, automotive navigation, medical health, gaming and entertainment, to improve end-user experience.