Huawei has made its MindSpore AI framework open source. With a large number of advantages that ‘AI-algorithms as-a-code’ can provide, the Chinese tech giant competes with the well-known AI frameworks of Google and Facebook.
In a statement, the Chinese tech giant states that its MindSpore AI framework is suitable for developing AI applications. The AI framework – co-developed with universities in Beijing and the United Kingdom, in addition to a Turkish startup – can easily be rolled out in various environments, such as on devices, within (multi)cloud and edge environments.
Important features of the framework include that it uses up to 20 percent less code than competing AI frameworks for different processes. Examples include neural language processsing (NLP).
MindSpore also supports simultaneous training on hardware to save training time. This would also enable dynamic debugging. This allows developers to more easily isolate bugs in their applications.
Furthermore, the AI framework of the Chinese tech giant has the advantage that it does not process data itself, but instead records the ‘gradient’ and model information that has already been processed. This ensures the integrity of sensitive data, according to Huawei.
According to the Chinese tech giant, MindSpore adapts native to all scenarios. This takes place by implementing ‘AI algorithms-as-a-code’ via on-demand collaboration. This makes it easier to roll out AI models. In addition, this approach helps improve collaboration with the Huawei AI Ascend processors. This hardware from the tech giant can significantly improve runtime efficiency and computing performance.
Hardware and software
The AI framework runs on various types of hardware, including processors, graphics cards and special neural processing chipsets. For these latter chipsets, think for example of Huawei’s Ascend applications.
For code, MindSpore now only supports Python 3.7+. However, the Chinese tech giant has plans to introduce support for C++, Julia and Rust in the near future. The framework runs most effectively on Linux platforms such as EulerOS and Ubuntu. Huawei MindSpore is now available via GitHub and Gitee.
Other AI Modules
In addition to Mindspore, the MindInsight open source module has also been created. This module helps developers with more debugging and tuning possibilities. This is done by making visualizations of the AI training process.
Visualizations include model parameter information such as training data and accuracy, measurement data of training progress and calculation graphs. A second module, MindArmour, helps to improve the security of the models to be developed.
Competition with US AI frameworks
By making his AI framework open source, Huawei appears to be consciously competing with similar open-source frameworks from tech giants Google (TensorFlow), and Facebook, (PyTorch), from the United States.
According to experts, one of the reasons for this is that the Chinese tech giant is now cut off from Google’s open source mobile OS Android. That is why Huawei would like to focus more on AI and acquire an important open source position there.