Pangu Large Model 3.0 is officially released. AI development is moving towards a new "industrialized development model"

Artificial intelligence is surging and is becoming the decisive force pushing mankind into the era of digital intelligence. On the afternoon of July 7, Huawei Managing Director and Huawei Cloud CEO Zhang Pingan announced at the Huawei Developer Conference 2023 that Huawei Cloud Pangu Large Model 3.0 was officially released. Zhang Pingan claimed that the architecture of Pangu Large Model 3.0 is divided into three levels: L0 basic large model, L1 industry large model and L2 scene model.

According to Zhang Ping'an's introduction, the L0 layer consists of five basic large models of natural language, vision, multi-modality, prediction and scientific computing. The L1 layer is N large industry models. It can provide industry-wide large models trained using industry public data, including large models for government affairs, finance, manufacturing, mining, meteorology, etc.; it can also be based on the industry's own data, in the Pangu large model On the L0 and L1 layers, we train our own proprietary large models for customers. In contrast, the L2 layer can provide models for more segmented scenarios by providing "out-of-the-box" model services.

In short, unlike Baidu’s Wen Xin Yi Yan and Alibaba’s Tong Yi Qian Wen, the Pangu model emphasizes the application in subdivided scenarios and mainly solves the problem of low-cost mass customization in the business environment. Zhang Ping'an also emphasized that Pangu's large model is a large model that is completely industry-oriented, including NLP large model, multi-modal large model, visual large model, prediction large model and scientific computing large model, which can be applied in various vertical scenarios.

At the Chengdu branch, in the face of a new round of technological and industrial revolution, how Chengdu can stand at the forefront of the development wave of the artificial intelligence industry, Huawei Cloud Sichuan Hidden Solution Minister Hu Chenglin also officially released the Pangu model Sichuan Ecological policies, including start-up support programs, specialization and innovation acceleration programs, and many other measures are used to stimulate the innovation vitality of Sichuan enterprises and help them develop innovatively.

During the conference, the Sichuan Division of the 2023 Shengteng AI Innovation Competition was also officially announced. The relevant person in charge of the Chengdu Intelligent Computing Center said that it will continue to strengthen the pace of local talent training and scientific research and innovation. In the critical acceleration period of scientific research breakthroughs and industrial application implementation in the artificial intelligence industry, in order to continue to explore the frontiers of industrial development, cultivate and develop AI talents for Sichuan, and promote the transformation of scientific research results.

Artificial intelligence drives the intelligent development of thousands of industries. AI is entering the core production system of enterprises, creating greater value in fields such as smart parks, smart medical care, digital government affairs, and smart factories. Large-scale models act as a bridge between technology and business ecology, can continuously provide intelligent upgrades for various industries, and promote high-quality development. There is a basic consensus in the industry that in the future, each industry will have its own vertical large model. The focus is on how to integrate the capabilities of the general large model with the company's own industry data.

The "China Artificial Intelligence Large Model Map Research Report" released by the New Generation Artificial Intelligence Development Research Center of the Ministry of Science and Technology shows that the number of large models developed by China ranks second in the world, second only to the United States. Global competition in the field of artificial intelligence large models has become fierce. In the past six months, major manufacturers such as Huawei, Baidu, and Alibaba have invested a lot of resources in the research and development of large models.

At this World Artificial Intelligence Conference, large models are also the key words throughout. IDC predicts that China's artificial intelligence software and application market will reach US$21.1 billion in 2026, and artificial intelligence has entered a critical period for large-scale implementation. Industry experts believe that "in-depth practice" is the key to the development of artificial intelligence, which can empower the industry and promote upgrading.

The above is the detailed content of Pangu Large Model 3.0 is officially released. AI development is moving towards a new "industrialized development model". For more information, please follow other related articles on the PHP Chinese website!