摘要
ChatGPT以其强大的自然语言处理和知识推理技术,在全球范围内引起了大型语言模型(LLM)的热潮。通过基于多租户基础架构的企业GPT构造器,实现了企业网站抓取和文档加载功能,为中小型企业租户提供了私域数据喂养和向量化服务,并将语料向量作为LLM的长期记忆,构建了企业级GPT能力。通过配置的5G消息Chatbot同企业客户进行交互,在交互过程中,其能通过检索向量数据库中的企业相关信息,提供企业级的ChatGPT服务。
With its powerful natural language processing and knowledge reasoning capability,ChatGPT has caused the craze of large language models(LLM)around the world.Through the enterprise GPT builder based on multi-tenant infrastructure,it realizes the functions of enterprise website crawling and document loading,provides private domain data feeding and embedding service for small and medium sized enterprise tenants,and uses embedded vectors as the long term memory of LLM to build enterprise level GPT capabilities.Through the configured 5G message Chatbot to interact with enterprise customers,it can provide enterprise level ChatGPT services by retrieving enterprise-related information in the vector database.
作者
李红纲
丁健
LI Honggang;DING Jian(Whale Cloud Technology Ltd,Nanjing 210012,China)
出处
《移动信息》
2024年第3期280-282,共3页
MOBILE INFORMATION
