Neural network structure searching method, device and equipment

Abstract

The embodiment of the invention discloses a neural network structure searching method, device and equipment. After a set of hyper-parameters output by the hyper-parameter search model is obtained, structural hyper-parameters in the set of hyper-parameters are used for searching to obtain a current neural network basic unit structure, so that a first neural network model is established according tothe current neural network basic unit structure; the first neural network model is trained by using the training hyper-parameters in the group of hyper-parameters; and a second neural network model is generated, and the second neural network model is tested to obtain an evaluation result, so as to update the hyper-parameter search model based on the evaluation result and return to execute the step of obtaining a group of hyper-parameters output by the hyper-parameter search model again; and the search is repeated until a preset stop condition is reached, and the current second neural networkmodel is determined as the target neural network model. Therefore, the task execution effect of the neural network model can be effectively improved.

Description

technical field [0001] The present application relates to the technical field of automatic machine learning, in particular to a neural network structure search method, device and equipment. Background technique [0002] The neural network model is the preferred solution for realizing artificial intelligence. The construction process of the general neural network model can include: (1) preprocessing the training verification data set (wherein, the preprocessing can include data enhancement, data balance, data normalization, etc. processing means); (2) build the network structure of the neural network model (that is, determine the number of nodes in the neural network model, node operation type, and node connection relationship and other structural information); (3) use the training verification data set to analyze the neural network model Perform training and verification to obtain the final neural network model. [0003] Among them, because the network structure of the neur...

AI Technical Summary

Problems solved by technology

However, because the neural network structure is determined by a large number of hyperparameters, and the process of determining a large number of hyperparameters is very difficult, the search process of the neural network structure is difficult.

Images