Deep neural network model compression method based on pruning threshold automatic search

Abstract

The invention discloses a deep neural network model compression method based on pruning threshold automatic search, and belongs to the field of deep neural network model compression. The method comprises the following steps: training a model to obtain an initial model for pruning; performing adaptive grid search on the model parameters to obtain a first pruning threshold; further searching for a threshold interval corresponding to the first pruning threshold by combining a binary search method, and searching for a better threshold to obtain a second pruning threshold; performing iterative pruning processing on the original network model based on the second pruning threshold; and carrying out sparse storage on the pruned model to obtain an available compressed network model. According to the deep neural network model compression method based on pruning threshold automatic search, an existing main deep neural network model can be compressed, the technical problem that the deep neural network model cannot be deployed to embedded equipment due to the fact that the model is large is solved, and the application range of the deep neural network model is expanded.

Description

technical field [0001] The invention belongs to the field of deep neural network model compression, in particular to a deep neural network model compression method based on pruning threshold automatic search. Background technique [0002] The development of deep learning has made deep neural networks more and more applied to computer vision tasks such as image recognition, detection and tracking, and more and more network models tend to be designed in a wider and deeper direction. The success of deep learning largely depends on a large number of parameters of the model and computing equipment with powerful performance. However, due to the huge memory requirements and computational consumption of deep neural networks, it is difficult to deploy them on low-storage, low-power hardware platforms (such as mobile devices), which greatly limits their applications. Therefore, it is an important issue to study how to effectively compress the neural network model while keeping the pe...

AI Technical Summary

Problems solved by technology

There are two problems here. One problem is that the definition of the pruning rate is artificially specified, rather than the model is automatically found. There may be a better value for the model pruning threshold; the other problem is that excessive pruning may make the model accuracy difficult The relationship between recovery, model accuracy and model compression rate is difficult to achieve a good balance

Images