爱数智慧人机交互AI训练数据集-方言

爱数智慧拥有全球最大的ASR数据库之一，涵盖中文，方言，重口音，欧洲小语种，东南亚小语种，阿拉伯等地区全地区覆盖，总量共计20W小时，标注准确率高，覆盖各类型应用场景，且全部有相关授权，目前已经为世界范围内各大头部科技公司提供过数据服务，包括阿里，腾讯，华为，微软，英伟达等企业

AISpeech owns one of the world's largest ASR databases, which covers comprehensive language resources including Mandarin Chinese, its dialects, heavily accented speech, minor European languages, minor Southeast Asian languages, Arabic and other regional languages with full regional coverage. The total duration of the database reaches 200,000 hours, boasting high annotation accuracy and supporting all types of application scenarios. All the data in the database has obtained relevant official authorizations, and the dataset has provided data services for leading global technology enterprises including Alibaba, Tencent, Huawei, Microsoft, NVIDIA and other enterprises.

搜集汇总

背景与挑战

背景概述

该方言数据集覆盖全球多种语言及方言，总量达20万小时，标注精准且场景全面，已获授权并为多家头部科技公司提供服务。