Neural network accelerator

Abstract

The invention discloses a neural network accelerator. The neural network accelerator is used for being externally connected with a global control module, a first direct memory access module DMA and amemory. The neural network accelerator comprises a second direct memory access module DMA, a convolution module, a single data processing module, a plane data processing module, a channel data processing module and a probability calculation module which are connected in sequence in an assembly line mode; the convolution module performs multiply-add operation on the data; the single data processingmodule performs normalization, proportional operation and activation function processing; the plane data processing module performs maximum pooling, minimum pooling and average pooling processing; the channel data processing module performs channel splicing, surface rearrangement and matrix replacement processing; the probability calculation module finds out five maximum values in the data and completes probability operation of the five maximum values; the second DMA transmits the data to a convolution module; and the convolution module and the channel data processing module share one DMA control bus.

Description

technical field [0001] The present invention relates to the technical field of integrated circuits, and more specifically, to a neural network accelerator. Background technique [0002] Convolutional Neural Network (CNN) is an important algorithm for deep learning, and it has a very wide range of applications in the field of computer vision, especially image recognition. At present, almost all identification and detection problems use convolutional neural network as the preferred method, and various IT giants in the world are also competing to carry out related research. [0003] From the perspective of the computer, the image is actually a two-dimensional matrix. The work of the convolutional neural network is to extract features from the two-dimensional array by using operations such as convolution and pooling, and to recognize the image. Theoretically, as long as the data can be converted into a two-dimensional matrix, it can be identified and detected by using a convolu...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem that the existing technology cannot make full use of the characteristics of the convolutional neural network, resulting in low computational efficiency of the neural network, a neural network accelerator is provided, which can improve the computational efficiency of the neural network and save computational time

Images