Yitu Technology marches into the voice field and unites Huawei Microsoft to push the voice open platform

category:Internet
 Yitu Technology marches into the voice field and unites Huawei Microsoft to push the voice open platform


According to Lv Hao, chief Innovation Officer of Yitu, the three test subsets of AISHELL-2 recently released in the industry, as well as the near-field accent test set, near-field quiet chat test set, voice program test set, telephone test set and far-field test set from third parties, are all at the leading level, and the error rate is almost 15%. Following. Among them, in AISHELL 2-2018A-EVAL data set, the accuracy of image recognition is as high as 96.29%, and the error rate (CER) is only 3.71%.

In addition, Yitu and Microsoft announced further cooperation and jointly launched Yitu Voice Open Platform. The platform will be based on Microsoft Azure cloud, open voice recognition technology capabilities to third-party application developers, and build AI ecosystem. At the same time, Yitu and Huawei jointly released the Intelligent Voice Joint Solution, which is based on Yitu Voice Open Platform, Huawei Ascend Series Chips and ATlas300AI Acceleration Card for Data Center Side. At the launch site, it is predicted that the performance of speech recognition algorithms will increase exponentially in the next 6 to 12 months, and more scenarios will be unlocked. (Dingxi) Focus on NetEase Smart Public Number (smartman 163) to interpret major corporate events in the field of AI for you, new ideas and new applications. Source: Netease Intelligent Responsible Editor: Wang Chao_NT4133

In addition, Yitu and Microsoft announced further cooperation and jointly launched Yitu Voice Open Platform. The platform will be based on Microsoft Azure cloud, open voice recognition technology capabilities to third-party application developers, and build AI ecosystem. At the same time, Yitu and Huawei jointly released the Intelligent Voice Joint Solution, which is based on Yitu Voice Open Platform, Huawei Ascend Series Chips and ATlas300AI Acceleration Card for Data Center Side.

At the launch site, it is predicted that the performance of speech recognition algorithms will increase exponentially in the next 6 to 12 months, and more scenarios will be unlocked. (Dingxi)

Focus on the smartman 163, to interpret the major events of AI companies, new ideas and new applications.