Home  >  Article  >  Technology peripherals  >  Omdia Research: PyTorch, TensorFlow, MindSpore, etc. are in the first echelon of China’s AI framework market

Omdia Research: PyTorch, TensorFlow, MindSpore, etc. are in the first echelon of China’s AI framework market

WBOY
WBOYforward
2023-04-12 19:40:011486browse

Omdia Research: PyTorch, TensorFlow, MindSpore, etc. are in the first echelon of China’s AI framework market

Industry research organization Omdia (a subsidiary of Informa tech group) released the "China Artificial Intelligence Framework Market Research Report", pointing out the competition pattern and innovation trends of China's AI framework market. Omdia research found that PyTorch, TensorFlow and MindSpore are in the first echelon in terms of popularity and market share.

Omdia Research: PyTorch, TensorFlow, MindSpore, etc. are in the first echelon of China’s AI framework market

With the development of technology, artificial intelligence has gradually appeared in front of the public. As an emerging industry, the market scale continues to expand, and application scenarios have also expanded with the maturity of artificial intelligence technology. From scientific research innovation to industrial application, pre-trained large models, AI for Science, responsible AI, etc. have become the focus of global academia and industry. The AI ​​framework is the core of model algorithm development and the basis for supporting the prosperity and development of the artificial industry. Therefore, Omdia has conducted an in-depth study of the development and industry trends of AI frameworks.

Nowadays, the attention of large models represented by ChatGPT has surpassed academia and industry, and has become an innovative achievement that everyone pays attention to, and large models require strong AI framework technology support. The report points out three major trends in pre-training large models: First, the parameter volume of large models continues to grow exponentially, and "large models" are moving towards "super-large models"; second, large models are moving from single-modal to multi-modal. Modal and multi-task fusion; Third, the artificial intelligence framework plays a key technical supporting role in the training of large models.

In terms of supporting the training and development of ultra-large-scale models, the world's leading artificial intelligence frameworks TensorFlow and PyTorch still occupy a leading position; at the same time, developers believe that MindSpore has taken a dominant position among China's local artificial intelligence frameworks, with native Support large models and incubate a series of innovative large models.

Through in-depth interviews with experts, Omdia found that TensorFlow has brought more expectations to the industry due to the integration of JAX, a new framework; PyTorch relies on third-party parallel algorithm libraries to supplement large model support capabilities; In the Chinese local market, Baidu Feipiao and MindSpore are more successful in supporting local pre-trained large models due to their unique local Chinese language and data advantages.

Behind the popularity of AIGC represented by ChatGPT, AI ethical and moral issues such as "counterfeiting" have also emerged. Artificial intelligence developers and institutions are paying more and more attention to "responsible artificial intelligence." In a survey of developers, Omdia found that among all mainstream artificial intelligence frameworks, TensorFlow and MindSpore provide the best support for "responsible artificial intelligence", ranking first and second respectively.

At the same time, the "China Artificial Intelligence Framework Market Research Report" pointed out that "responsible artificial intelligence" is both a set of moral principles and a set of technical systems. “Responsible AI” is an approach to developing, evaluating, deploying and scaling artificial intelligence systems in a safe, reliable and ethical manner. The artificial intelligence framework introduces numerous technical means and trusted AI functional modules to help developers create trusted AI, help developers and institutions solve compliance issues such as the security and privacy of artificial intelligence, and achieve sustainable development of artificial intelligence.

In the field of scientific research and innovation, "AI for Science" is also a cutting-edge hot spot in the artificial intelligence industry. The deep integration of artificial intelligence and science is promoting the innovation of scientific research paradigms and bringing new developments to the scientific research field. opportunity. Omdia analysts believe that, similar to large models, "AI for Science" is a new important direction for the development of AI innovation, and the artificial intelligence framework plays a key technical support role in the development of "AI for Science". The development of "AI for Science" has put forward higher requirements for artificial intelligence frameworks. The survey found that Chinese artificial intelligence developers believe that MindSpore is the most suitable domestic artificial intelligence framework for the "AI for Science" project. The support capability of "AI for Science" even exceeds that of PyTorch, and has a tendency to catch up with TensorFlow.

As an important foundation for AI innovation, the artificial intelligence framework will help the industry accelerate intelligent transformation and upgrading. For more research findings on artificial intelligence frameworks, please view the "China Artificial Intelligence Framework Market Research Report". Omdia report link:

​https://www.php.cn/link/6562a2c4889bc3e203ab869723a40f7e​

The above is the detailed content of Omdia Research: PyTorch, TensorFlow, MindSpore, etc. are in the first echelon of China’s AI framework market. For more information, please follow other related articles on the PHP Chinese website!

Statement:
This article is reproduced at:51cto.com. If there is any infringement, please contact admin@php.cn delete