Method and device for generating neural network, computer equipment and storage medium

Abstract

The invention provides a method and device for generating a neural network, computer equipment and a storage medium. The method comprises the steps that: a first search space corresponding to a preset network structure unit, a second search space corresponding to a down-sampling processing unit and a third search space corresponding to the size information of a neural network structure are determined, wherein the preset network structure unit and the down-sampling processing unit are used for constructing the neural network structure; based on the first search space, the second search space, the third search space and a training data set corresponding to a preset deep learning task, network structure search is performed to obtain a target neural network structure corresponding to an expected result of the deep learning task; and a target neural network for processing the deep learning task is generated based on the target neural network structure.

Description

technical field [0001] The present disclosure relates to the technical field of deep learning, in particular, to a method, device, computer equipment and storage medium for generating a neural network. Background technique [0002] For deep learning tasks, such as image classification tasks, image detection tasks, or image segmentation tasks, it often requires different types of network structures to jointly construct a deep neural network to perform. There are many choices for each network structure, and the combination of different types of network structures will produce neural networks with different structures. The performance of these neural networks in the corresponding deep learning tasks depends on the pros and cons of the network structure. Aiming at the combination of different network structures, nowadays it mainly relies on manually building neural networks and relying on human experience to build them. After training, the performance of the trained neural netwo...

AI Technical Summary

Problems solved by technology

For the combination of different network structures, nowadays, it mainly relies on manual construction of neural networks and artificial experience. In this way, the performance of the trained neural network in deep learning tasks such as accuracy and resource consumption needs to be improved.

Images