Algorithm model deployment online method and device

Abstract

The invention discloses an algorithm model deployment online method and device, and relates to the technical field of computers. A specific embodiment of the algorithm model deployment online method comprises the following steps: reviewing a to-be-deployed algorithm model, merging the algorithm model into a to-be-deployed service architecture after the algorithm model passes the reviewing, and generating a pre-sent version package; carrying out model verification on the pre-issued version package through a pre-issued test environment; and realizing online deployment of an algorithm model basedon the verified pre-sent version package. According to the implementation mode, the algorithm model and the service architecture are decoupled, and automatic agile online of the algorithm model can be realized, so that the algorithm model of the recommended prediction service is updated in time, and the online process is simple, and the online efficiency is high, and the risk is low.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a method and device for deploying and launching an algorithm model. Background technique [0002] With the development of Internet technology, personalized recommendation has become an important means of attracting user traffic. The current recommendation architecture has undergone a technological innovation—recommendation and prediction service items. The service can adjust the recommendation results in real time according to activities, users themselves, etc., and can set experimental positions to evaluate whether the algorithm model is accurate through a small amount of real traffic. In order to ensure the superiority of the prediction service, that is, to be able to launch more appropriate results, the algorithm model needs to be frequently launched for verification and tuning. [0003] When adjusting the algorithm model in the prior art, it is necessary to merge the algo...

AI Technical Summary

Problems solved by technology

[0005] (1) Unable to support frequent online, causing the algorithm model of the recommendation and prediction service to be unable to be updated in time;

[0006] (2) Manually go online through development, testing, operation and maintenance personnel, and the risk of going online is high;

[0007] (3) Every time it goes online, the algorithm model to be deployed needs to be merged into the forecast service architecture, and then the forecast service is launched as a whole. The online process is cumbersome and the online efficiency is low

Images