Shulou(Shulou.com)11/24 Report--

CTOnews.com news on April 6th, at present, with the rapid development of artificial intelligence, high-performance computing, 5G and other fields, digital, intelligent and diversified applications continue to emerge, which puts forward higher requirements for computing network infrastructure data processing and network transmission capacity.

As an important part of artificial intelligence engineering, artificial intelligence R & D operation system (MLOps) aims at building a standardized, integrated and systematic production and operation system for the whole life cycle of AI model, which promotes the transition of model production from decentralized small workshop model to large-scale and orderly factory assembly line model, and helps organizations to improve the landing quality and efficiency of AI.

In order to systematically sort out the components of the MLOps tool chain, gain a comprehensive insight into the current situation of industrial development, effectively assist the application side in technology selection, and continue to promote the high-quality development of the industry, the Chinese Academy of Information and Communications has started the compilation of "MLOps tool atlas".

In addition, in order to promote the efficient connection between supply and demand of industrial supply and demand, and to explore the capability application and product deployment of on-line computing in scenarios such as intelligent computing center and 5G private network, the China Institute of Information and Communications Research and Network Convergence Industry and Standards Promotion Committee (CCSA TC621) jointly launched the "INC Ready" project to carry out the benchmark capability assessment of online computing products and solutions, with the aim of building industry consensus. Form innovative technologies and key product R & D capabilities to promote the ecologically healthy development of the online computing industry.

INC Ready industry benchmark capability assessment divides the content of online computing products and solutions industry benchmark capability assessment into three categories: facilities, solutions and services, with a total of 9 modules, from the dimensions of technology development trend, product research and development status, application scenario requirements and other dimensions. The evaluation results will provide comprehensive reference for vertical industry users to choose products and solutions.

CTOnews.com learned officially from the China Academy of Information and Communication that the MLOps tool atlas can be divided into two categories:

1. End-to-end platform

During the whole life cycle of the model, the MLOps platform supports the whole process from data preparation to model launch. In addition, have a certain degree of asset management and basic security capabilities.

two。 Special tool

It is subdivided into three categories: process capability, product management and basic guarantee. A MLOps platform or tool that supports one or more fine-grained tasks (such as data processing, data quality inspection, data annotation, feature management, etc.) during the whole life cycle of the model.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.