Megvii unveils latest version of open-source deep learning framework for algorithm training
Published: Sep 18, 2020 06:03 PM

The AI fever screening system scans a visitor in Haidian District Service Hall, Beijing, Feb. 4, 2020. (Photo: Xinhua)

Chinese artificial intelligence (AI) firm Megvii Technology, known for its facial recognition platform Face++, on Friday unveiled the 1.0 preview version of its proprietary open-source deep learning framework MegEngine, in a fresh move to close China's gap in AI ecosystem with the US.

At a meeting on AI, open source and productivity during the 2020 Zhongguancun Forum, a sci-tech forum in Beijing, Sun Jian, chief scientist of Megvii and head of Megvii Research, took the wraps off the latest version of MegEngine, part of the firm's AI architecture Brain++ that also includes a data management system and a computer power dispatching platform.

The Beijing-based AI unicorn announced in March to open-source MegEngine, making it accessible to developers from across the globe. 

From the availability of the alpha version in March to the launch of the 1.0 preview version, the self-developed deep learning framework has gone through eight iterations, upgrading to 510,000 lines of code from the previous 350,000 lines of code, according to Sun. Currently, all of Megvii's algorithms are trained and inferred through MegEngine.

The latest version features technology upgrades including automatic code stripping, inference performance improvements and enhanced support for homemade hardware.

Megvii's upping the ante in AI infrastructure is seen to reduce the nation's dependence on US-origin deep learning frameworks, especially as concerns mount over a technology split.  

Megvii and two other major AI start-ups SenseTime Group and YITU Technology were among eight Chinese tech firms that were added to the US Entity List in October 2019. 

There is still a long way off for China's proprietary AI frameworks to catch up with their more popular US counterparts, according to Yuan Jinhui, founder of OneFlow Inc, a Beijing-based firm specializing in developing AI infrastructure. Google's TensorFlow and Facebook-backed PyTorch take the lion's share of AI frameworks globally. 

"For an AI ecosystem to rise to prominence, such frameworks need to pack unique features to attract developers, which is essentially about being innovative to sharpen their competitiveness," Yuan told the Global Times on Friday.

Multiple Chinese alternatives to US heavyweights that have been up and running have already proven to be unique in certain aspects from existing open-source frameworks in that they are much faster or support functionalities that are yet to be available on other frameworks, he commented. 

OneFlow also open-sourced its self-developed deep learning framework earlier this year.