Wang Haifeng, CTO of Baidu: Flying oar ranks first in the comprehensive market share of China's deep learning platform

TechWeb 2022-01-15 03:13:04

【TechWeb】12 month 11 Daily news , stay WAVE SUMMIT+ 2021 In depth learning developer Summit , Baidu CTO Wang Haifeng announced the transcript of the open source deep learning platform flying oar : condensation 406 Million developers , establish 47.6 Wan model , service 15.7 Ten thousand enterprises and institutions , The comprehensive market share of China's deep learning platform ranks first .

Baidu CTO Wang Haifeng

Wang Haifeng said , AI enters the stage of industrial mass production , We need to AI Mass production platform , bring AI Technology can be standardized 、 Output to all walks of life in an automated and modular manner , Realize large-scale application ; At the same time, it can unite all parties , Platform based integration and innovation , common development . The open source open platform for deep learning of the propeller industry is a typical example AI Mass production platform , Empowering developers , Strong support AI Industrial mass production , Promote technological innovation and industrial intelligence upgrading .

Wang Haifeng believes that , Artificial intelligence presents the characteristics of integrating innovation and reducing the threshold : One side ,AI The integration and innovation of technology and industry are more and more ; On the other hand , although AI Technology is getting more and more complex , but AI The threshold of development and application is getting lower and lower . With AI Deep integration with the industry , In the future, more and more complex AI personnel , Know both AI, And industrial experience .

At this summit , Flying oars release ten latest technological and ecological advances , Including the new panorama of the propeller — New Wenxin big model in industrial model library 、 The industry's first industrial practice example library 、 Flying propeller “ A great voyage ”2.0 Co creation plan 、 Flying oar open source framework v2.2 Systematic new scientific computing API、 End to end adaptive large-scale distributed training technology 、 The whole process of text task is accelerated 、 Multi level and low-cost hardware adaptation scheme 、 The industrial open source model library model exceeds 400 individual 、 The enterprise edition upgrades the automatic and efficient model deployment function , as well as 1 Minutes of high-speed installation to complete the local efficient modeling of the propeller EasyDL The desktop version .