Neural network model search method, image recognition method and device

Abstract

The present application relates to a neural network model searching method, image recognition method, apparatus, computer equipment and readable storage medium. The method includes: acquiring training sample data, test sample data and a plurality of preset initial neural network models with different structures; according to the training sample data, a plurality of initial neural network models and a plurality of preset network layers with different structures, derive A plurality of neural network models are generated, and at least one neural network model with an optimal test result is selected from the plurality of neural network models as the target neural network model according to the test results of the test sample data. The method can improve the design efficiency of the neural network model.

Description

technical field [0001] The present application relates to the technical field of neural networks, and in particular, to a neural network model search method, image recognition method, apparatus, computer equipment and readable storage medium. Background technique [0002] Neural network models have been widely used in image processing, speech processing, text processing and other fields. Taking the most commonly used convolutional neural network as an example, network structures with better performance have been proposed in recent years, including but not limited to AlexNet, VGG16, Inception, Resnet, Xception, etc. [0003] For a specific task or data set, researchers can continuously propose convolutional neural network models with matching performance through manual design and experimentation. However, the manual design method requires a lot of researcher's ability and experience. If there is a lack of relevant theoretical guidance, it often takes researchers a lot of tim...

AI Technical Summary

Problems solved by technology

However, the manual design method has high requirements on the ability and experience of researchers. If there is no relevant theoretical guidance, it often takes a lot of time and energy for researchers to design an applicable convolutional neural network for a certain task or a certain data set. Network model; moreover, there are many ways to design the model, and the number of model structures considered by the manual design method is very limited, and the designed model is often quite different from the optimal model in terms of accuracy, and there is a lot of room for improvement

[0004] In short, the traditional way of manually designing the neural network model has limitations such as low design efficiency, which makes it difficult to achieve the optimal performance of the designed neural network model.

Images