Convolutional neural network hardware module deployment method

Abstract

The invention discloses a convolutional neural network hardware module deployment method, and relates to the field of convolutional neural network implementation. The method comprises the steps that an upper layer compiler of a convolutional neural network is used to perform simulation comparison on the implementation form of the convolutional neural network according to a target hardware resourceand a data volume of a convolutional neural network model; the requirements of the hardware resource and convolutional neural network speed are weighed to determine a deployment parameter of each hardware module of the convolutional neural network; the number of each hardware module is divided, and the connection mode among hardware modules is determined, so that the deployment of the convolutional neural network hardware modules is achieved.

Description

technical field [0001] The invention discloses a convolutional neural network hardware module deployment method, and relates to the field of convolutional neural network implementation. Background technique [0002] Convolutional neural network (CNN) is a variant model of multi-layer perceptron (MLP). It evolved from the biological concept that the visual cortex cells cover the entire visual domain in a certain way, like some filters, they are locally sensitive to the input image, so they can better mine natural images. The spatial relationship information of the target. CNN mines the spatial local association information of objects of interest in natural images by strengthening the local connection patterns of nodes between adjacent layers in the neural network. [0003] Most of CNN's deployment methods currently use the X86 architecture CPU platform + GPU as the hardware environment, run software frameworks such as TensorFlow, Caffe, etc. in the real-time operating syste...

AI Technical Summary

Problems solved by technology

On the application side, due to resource constraints, it is often only possible to run a small model, and the model operation efficiency is low, and the recognition speed is much lower than that of the PC side.

Images