Huawei Cloud ModelArts Continues to Rank No. 1 in China's Public Cloud Service Market for Machine Learning
Dec 24, 2021
IDC recently released "2021H1 China AI Cloud Service Market Research Report". According to the report, Huawei Cloud's one-stop AI development platform ModelArts ranked No. 1 in China's public cloud service market for machine learning, with a market share of 30.6%, a year-over-year growth of 1.6%. ModelArts has now topped the market for three consecutive years.
According to the report, Huawei Cloud is favored by government and enterprise customers in China's public cloud service market for machine learning, demonstrating its increasingly competitive advantages in the AI cloud market.
ModelArts provides full-lifecyle AI development including data processing, algorithm development, model training, model management, and model deployment. ModelArts is the first to support MLOps and continuously builds large model training and optimization capabilities. Moreover, Huawei introduced ecosystem development tools to the market and offers an open AI ecosystem. Huawei Cloud AI Gallery connects the supply, demand, and learning sides of the AI market to solve the problems of insufficient AI computing power, AI talent shortages, and difficult AI development and implementation, enabling AI to benefit various industries.
Providing cutting-edge technologies for AI developers
ModelArts supports large-scale distributed training. To help developers and enterprise customers address the technical, cost, and resource problems in large model training, ModelArts provides AI cluster services, which are optimized in multiple scenarios for large-scale distributed training. ModelArts offers four times higher development efficiency compared with on-premises large model development and solves the problem of resource availability for ultra-large model training.
The cloud-native model development tool chain improves AI development efficiency. ModelArts combines on-cloud resources with the development tool chain. For AI beginners, the on-cloud CodeLab enables resource access within seconds and provides on-demand switchover. For AI professionals, ModelArts allows them to customize development environments and remotely develop models on the cloud. Developers can use ModelArts cloud-native resources to further build algorithm development paradigms and capabilities.
Building an AI development community featuring "knowledge" and "practice" for the AI ecosystem
Based on ModelArts, Huawei Cloud has built AI Gallery, a developer ecosystem community. It is a social platform for AI asset sharing and also an AI development community featuring "knowledge" and "practice". AI Gallery provides more than 50,000 AI assets of more than 10 types, including algorithms, models, datasets, and workflows, to enable all participants in the AI development and application ecosystem to efficiently achieve business values, reduce difficulty in learning AI, and accelerate AI implementation.
Providing end-to-end MLOps support for AI implementation
The original phase-by-phase AI model development practice is replaced by end-to-end, automated, collaborative, and iterative development. ModelArts is now the first AI development platform with end-to-end MLOps support in China.
Based on MLOps, ModelArts workflows provide functions such as running recording, monitoring, and continuous running. The workflow toolkit enables lifecycle management that covers AI development, running, and monitoring. It accelerates application iteration from development to implementation, and allows for continuous optimization of application performance. In addition, developers can use AI Gallery to quickly build and share workflows.
ModelArts provides enterprises with AI lifecycle management, greatly improving AI development efficiency and accelerating AI implementation in various industries.
In the manufacturing field, Bosch and Huawei Cloud have developed a use case for intelligently detecting tool statuses based on ModelArts workflows. With the sound recognition algorithm, tool usage efficiency improved by more than 5% with a stable yield rate. The concept behind the use case for intelligently detecting tool statuses can be easily applied to different tools and even other production tools, and the sound recognition algorithm has been used for tool status maintenance in a wide range of areas.
In the drug R&D field, working with Huawei Cloud ModelArts and EIHealth, the Shanghai Institute of Materia Medica, Chinese Academy of Sciences has developed the Huawei Cloud Pangu drug molecule model, enabling end-to-end AI drug design and greatly improving the efficiency of drug R&D.
In the transportation field, Shenzhen Airport has quickly built more than 60 scheduling rules based on Huawei Cloud Airport Intelligent Twins for automatic and intelligent stand allocation, improving the aircraft docking rate by 5% and freeing 2.6 million passengers each year from having to take ferry buses.
In the future, Huawei Cloud will continue to innovate, build an open and collaborative Huawei Cloud global ecosystem, and make technologies available as cloud services to customers in various industries. In this way, Huawei Cloud aims to provide a ubiquitous cloud foundation for an intelligent world.