信息通信技术与政策

信息通信技术与政策

信息通信技术与政策 ›› 2024, Vol. 50 ›› Issue (6): 45-53.doi: 10.12267/j.issn.2096-5931.2024.06.007

专题:先进计算 上一篇    下一篇

一种基于大模型技术的算力产业监测服务平台设计

A design of monitoring service platform for computing power industry based on large language model technology

李峰1, 乔春庚2   

  1. 1.中国信息通信研究院信息化与工业化融合研究所,北京 100191
    2.拓尔思信息技术股份有限公司,北京 100096
  • 收稿日期:2024-05-15 出版日期:2024-06-25 发布日期:2024-07-30
  • 作者简介:
    李峰, 中国信息通信研究院信息化与工业化融合研究所工程师,主要从事人工智能芯片评测和大模型应用等方面的研究工作;
    乔春庚, 拓尔思信息技术股份有限公司工程师,主要从事自然语言处理和大模型产品研发等方面的工作

LI Feng1, QIAO Chungeng2   

  1. 1. Informatization and Industrialization Integration Research Institute, China Academy of Information and Communications Technology, Beijing 100191, China
    2. TRS Information Technology Co., Ltd., Beijing 100096, China
  • Received:2024-05-15 Online:2024-06-25 Published:2024-07-30

摘要:

提出了一种基于大模型技术的算力产业监测服务平台的设计,旨在通过高效的技术手段实现对算力产业发展监测,为智能化决策提供更加综合全面的支撑。该平台采用分层架构设计,底层以国产百亿参数大模型基座为支撑,上层为业务层应用系统,提供算力发展洞察、智能分析助手等功能。

关键词: 大模型, 向量库, 算力产业监测

Abstract:

This paper proposes a design of monitoring service platform for computing power industry based on large language model technology, aiming to provide comprehensive support for intelligent decision-making in monitoring the development of the computing power industry by efficient technical means. The platform adopts a layered architecture design, with a bottom layer supported by a domestic large model base with 10 billion parameters, while the upper layer serves as a business application system, providing computing power development insights, intelligent analysis assistance, and other services.

Key words: large language model, vector database, computing power industry monitoring

中图分类号: