A method and device for generating a machine learning model

Abstract

The embodiment of the present application provides a method and device for generating a machine learning model, which is applied to a machine learning platform. The method includes: acquiring the machine learning requirements of the target server, the machine learning requirements including the training data format of the machine learning model and the machine learning The target function of the model; according to the machine learning requirements, use the preset model generation strategy to configure the target server for the target model of machine learning; return the target model to the target server. Therefore, data upload operations are simplified, network traffic is saved, the threshold for using machine learning models is lowered through modular settings, and the generation efficiency of machine learning models is improved.

Description

Technical field [0001] Technical Field The present application relates to machine learning, and more particularly to a machine learning model generation method, apparatus. Background technique [0002] Machine learning refers to past experience or data, to explore the laws of adaptation data so as to achieve self-learning. An existing machine learning machine learning internet platform is built on cloud platforms, the user submits data to the cloud platform, through a series of flow diagrams processed by the machine learning platform boxes, finally running out of the machine learning result. Therefore, the prior art there is a problem, the user's machine learning data is typically stored locally, you need to upload data to the cloud platform, the huge amount of data to be analyzed, the upload process is very time consuming. Meanwhile, when the target users of the system to be deployed not use cloud server machine learning platform correspondence, such as machine learning platform...

AI Technical Summary

Problems solved by technology

Therefore, the problem in the existing technology is that the user's machine learning data is generally stored locally, and the data needs to be uploaded to the cloud platform first. If the amount of data to be analyzed is huge, the upload process is very time-consuming

At the same time, when the target system to be deployed by the user does not use the cloud server corresponding to the machine learning platform, for example, the Alibaba machine learning platform corresponds to the Alibaba Cloud server, and the deployment of the target system can only be realized by uploading data to the corresponding cloud server, affecting System real-time

Images