Pipeline-based neural network processing system and processing method

Abstract

The present invention provides a neural network processing system. The processing system comprises a multiplier module, wherein the multiplier module contains a multi-level structure constituting a pipeline, and is configured to perform a multiplication operation of to-be-calculated neurons and weights in a neural network, and each level structure of the multiplier module completes the sub-operation of the multiplication of the neuron and the weight; an accumulator module, wherein the accumulator module comprises a multi-level structure constituting the pipeline, and is configured to perform an accumulation operation on the multiplication result of the multiplier module to obtain the output neurons of the convolutional layer in the neural network, and each level structure of the accumulator module completes the sub-operations of the accumulation operation; a pooling unit, wherein the pooling unit is configured to perform the pooling processing on the output neurons of the convolutionallayer; and a control unit, wherein the control unit is configured to control the transfer of data in the neural network processing system. By using the neural network processing system of the presentinvention, resource utilization and the data processing speed can be improved.

Description

technical field [0001] The invention relates to the technical field of deep learning, in particular to a pipeline-based neural network processing system and processing method. Background technique [0002] In recent years, deep learning technology has developed rapidly and has been widely used in solving high-level abstract cognitive problems, such as image recognition, speech recognition, natural language understanding, weather prediction, gene expression, content recommendation and intelligent robots. Research hotspots in academia and industry. [0003] Deep neural network is one of the perception models with the highest level of development in the field of artificial intelligence. It simulates the neural connection structure of the human brain by establishing a model, and describes the data characteristics hierarchically through multiple transformation stages, providing images, videos, audios, etc. Large-scale data processing tasks bring breakthrough progress. A deep ne...

AI Technical Summary

Problems solved by technology

[0004] However, in the prior art, the neural network has problems such as slow processing speed and high power consumption.

Images