Convolutional neural network model compression method, apparatus and device, and storage medium

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a convolutional neural network model compression method and device, equipment and a storage medium. The method comprises the steps of copying an original convolutional neural network model in an application to obtain N alternative models Mi; and compressing and training any two layers of convolution kernels of each alternative model Mi to obtain an adjusted alternative model Mi, and selecting an optimal alternative model Mk with minimum performance loss to run the application to obtain current internal environment parameters of the mobile terminal, taking the optimal alternative model Mk meeting a preset resource condition as a compressed convolutional neural network model; and otherwise, taking the optimal alternative model Mk as the original convolutional neural network model of the next round of model compression, and performing compression again. The invention also relates to a blockchain technology. The original convolutional neural network model is stored in a blockchain. According to the invention, the convolutional neural network model is automatically adapted to the mobile terminal for compression.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a convolutional neural network model compression method, device, equipment and storage medium. Background technique [0002] In the era of mobile Internet, people rely more and more on mobile devices such as mobile phones and tablets to obtain, use and generate information. Especially after the popularization of 4G and 5G networks, people's demand for digital multimedia information on the mobile terminal has also changed from pure text to Gradually evolving to images and videos, more and more people use mobile devices to process image and video data. The convolutional neural network model (Convolutional Neural Network, CNN) has developed into one of the most advanced technologies for computer vision tasks, which facilitates the processing of images and videos on mobile devices. [0003] At present, the main method to achieve lightweight models is still to...

AI Technical Summary

Problems solved by technology

[0004] The main purpose of the present invention is to solve the technical problem that existing convolutional neural network model compression methods cannot automatically adapt to mobile terminals

Images