Involving AI! Huawei’s major self-developed technology is officially open source
On the second day of Huawei Developer Conference 200 (Cloud), Huawei MindSpore Chief Scientist and IEEE Fellow Professor Chen Lei announced that MindSpore, Huawei’s all-scenario AI computing framework, is officially open source in Code Cloud, and will be committed to building a global open source community and continuing to promote AI The open source ecosystem of software and hardware applications flourishes.
In August last year, Huawei released the most powerful AI chip Shengteng 910 at its Shenzhen headquarters, and launched MindSpore, a full-scenario AI computing framework. At that time, Huawei’s rotating chairman Xu Zhijun said that in order to promote AI applications, Huawei’s full-scenario AI computing framework MindSpore will be open sourced in the first quarter of 2020.
MindSpore focuses on improving ease of use and lowering the development threshold for AI developers. Chen Lei said that MindSpore natively adapts to each scenario including end, edge and cloud, and can make the development state more friendly by implementing AI algorithms as code on the basis of on-demand collaboration, significantly reducing model development time and model development threshold.
Through MindSpore’s own technological innovation and the collaborative optimization of MindSpore and Huawei’s Ascend AI processor, the high efficiency of the running state has been achieved, and the computing performance has been greatly improved; MindSpore also supports other processors such as GPU and CPU.
In addition, for industry AI application developers, Yuan Jing, general manager of HUAWEI CLOUD General AI Services, released ModelArts Pro, the industry’s first professional enterprise-level AI application development suite. Huawei Cloud ModelArts Pro is positioned as an enterprise AI productivity tool, which can comprehensively improve the industry’s AI development efficiency and landing effect.
Huawei also demonstrated the full-scene development practices of various industry partners based on Huawei’s artificial intelligence computing platform Atlas. So far, the full-stack, full-scenario AI solution released by Huawei at the 2018 Full Connect Conference has been fully implemented for developers.
At the 2018 Huawei Full Connect Conference, Huawei released a full-stack full-scenario AI solution for the first time, including chips, chip enablement, training and inference frameworks, and application-enabled full-stack solutions, which can be used in public clouds, private clouds, and various Full-scene deployments such as edge computing, IoT industry terminals, and consumer terminals.
At the Shengteng 910 conference last year, when talking about Huawei’s AI strategy, Xu Zhijun said that Huawei positioned AI as a new general purpose technology, just like railways and electricity in the 19th century, and automobiles, computers, and the Internet in the 20th century. Will be applied to almost all places in the economy.
At the same time, Huawei also believes that the application of AI is still in the early stages of development, and AI technology and capabilities are still far behind long-term expectations. Reducing or even eliminating these gaps and accelerating the application of AI are the original intentions and goals of Huawei’s AI strategy.