Neural network structure searching and model publishing method, electronic equipment and storage medium

Abstract

The invention provides a neural network structure searching method, a model publishing method, electronic equipment and a storage medium. The method comprises the following steps: acquiring first information of a first application platform, wherein the first information is used for indicating a first computing power value of the first application platform; acquiring second information, wherein the second information is used for indicating performance parameters of a second application platform when the models with different calculated amounts are operated; converting the second information according to the first information of the first application platform based on a preset conversion rule to obtain third information; obtaining a target performance parameter of the target model, and determining the calculation amount of the target model in the third information according to the target performance parameter; according to the calculated amount of the target model, searching in a preset super network to obtain the target model; when the obtained target model is applied to the first application platform, the computing power of the first application platform can be fully utilized, and the computing power of the first application platform can ensure that the target model has relatively high operation efficiency.

Description

technical field [0001] The present application relates to the technical field of artificial intelligence, and in particular to a neural network structure search, a model release method, electronic equipment and a storage medium. Background technique [0002] The deep learning model has achieved very good results in many tasks, but the design of the neural network structure is very dependent on expert experience, the time period for manually designing the network structure is long, and the cost of hiring corresponding experts is high. The neural network structure search algorithm allows machines to replace experts in the design of neural network structures, and greatly improves model performance. The search efficiency is higher than that of manual design. [0003] At present, the artificially designed network model or the model obtained by the neural network structure search algorithm often cannot make full use of the computing power of the hardware platform or exceed the com...

AI Technical Summary

Problems solved by technology

[0003] At present, the artificially designed network model or the model obtained by the neural network structure search algorithm often cannot make full use of the computing power of the hardware platform or exceed the computing power of the hardware platform. When the computing power of the hardware platform cannot be fully utilized, the size of the model will be limited to the size of the model. Accuracy, and when the computing power of the hardware platform is exceeded, the operating efficiency of the model cannot meet the demand

Images