A neural network model training method, device and electronic equipment

Abstract

The invention relates to the field of pattern recognition, and discloses a neural network model training method, device and electronic equipment to solve the technical problem of low training stability in the prior art when training a neural network model through an asynchronous update algorithm of multiple computing devices . The method includes: in the process of training the neural network model, if the number of training rounds of training of at least two first computing devices meets a preset condition, based on the weight value of the main model of the neural network model, each second The weight value of the copy of the master model corresponding to a computing device is updated synchronously, so that the synchronization strategy of the weight value of the neural network model can be increased without significantly increasing the training time, ensuring the master model and master model. The consistency of the weight value of each copy achieves the technical effect of increasing the stability of the neural network model training.

Description

technical field [0001] The invention relates to the field of pattern recognition, in particular to a neural network model training method, device and electronic equipment. Background technique [0002] The neural network (NN: Neural Networks) model is a complex network system formed by a large number of simple processing units (called neurons) that are widely connected to each other. It reflects many basic characteristics of human brain functions and is a highly complex network system. nonlinear dynamic learning system. The neural network model has large-scale parallelism, distributed storage and processing, self-organization, self-adaptation and self-learning capabilities, and is especially suitable for dealing with imprecise and fuzzy information processing problems that need to consider many factors and conditions at the same time. [0003] In the prior art, the weight value of the neural network model is often adjusted through the learning and training process, and fina...

AI Technical Summary

Problems solved by technology

[0004] The invention provides a neural network model training method, device and electronic equipment to solve the technical problem of low training stability in the prior art when the neural network model is trained through the asynchronous update algorithm of multiple computing devices

Images