Model reasoning service calling system and method

Abstract

The embodiment of the invention discloses a model reasoning service calling system. The system comprises two business application clusters respectively supporting a local mode and a remote mode, each business application cluster comprises a plurality of business application nodes, each business application node is provided with an application container, and each application container is provided with a model reasoning service SDK (Software Development Kit); the system further comprises a reasoning service cluster which comprises a plurality of reasoning service containers, and each reasoning service container calls the model loaded in a model library for reasoning and returns a reasoning result; the system further comprises a feature library which is used for storing feature data of the models of the multiple versions, the model library which is used for storing other data of the models of the multiple versions, and an automatic publishing and deploying platform which is used for selecting the model from the model library according to a user request, and configuring the business application clusters and the reasoning service cluster. The embodiment of the invention further discloses a model reasoning service calling method. According to the invention, the algorithm model reasoning service can be efficiently completed, the algorithm model reasoning service and the business application are decoupled, and better double-track iteration can be realized.

Description

technical field [0001] The present invention relates to the field of computer technology, in particular to a system and method for invoking model reasoning services. Background technique [0002] In the field of AI, algorithm model reasoning is an important link. In the field of existing algorithm model reasoning services, there is no standardized process for algorithm model reasoning services, resulting in low iteration efficiency, and the coupling between business applications and algorithm model reasoning services is too tight, resulting in too much coupling between algorithm model training and iteration and system architecture evolution. Tight, mutual influence, difficult to update, low efficiency. Contents of the invention [0003] In order to solve the above problems, the purpose of the present invention is to provide a system and method for invoking model reasoning services, which can efficiently complete algorithm model reasoning services, and decouple algorithm m...

AI Technical Summary

Problems solved by technology

In the field of existing algorithm model reasoning services, there is no standardized process for algorithm model reasoning services, resulting in low iteration efficiency, and the coupling between business applications and algorithm model reasoning services is too tight, resulting in too much coupling between algorithm model training and iteration and system architecture evolution. Tight, interacting, hard to update, inefficient

Images