Neural network processor based on weight compression, design method, and chip

Abstract

The invention brings forward a neural network processor based on weight compression, a design method, and a chip. The processor comprises at least one storage unit used for storing operation instructions and data participating in calculation; at least one storage unit controller used for controlling the storage unit; at least one calculation unit used for executing calculation operation of a neural network; a control unit connected with the storage unit controller and the calculation unit for obtaining the instructions stored by the storage unit via the storage unit controller and analyzing the instructions so as to control the calculation unit; and at least one weight retrieval unit used for retrieving weights, wherein each weight retrieval unit is connected with the calculation unit so as to ensure correct operation on the compressed weights and the corresponding data. According to the invention, less weight resources in a neural network processor are occupied, the operation speed is improved, and the energy efficiency is improved.

Description

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine

Login to View More

AI Technical Summary

Benefits of technology

The technical effect of this patented method described in the previous figure was improved performance for Neural Networks (NN) due to its ability to efficiently process data without slowdown or consume too much power during training time. This compressed form also helps reduce resource usage on other parts within the system.

Problems solved by technology

This patented technical problem addressed in this patents relates to improving the computation speed and reducing resource usage required in deep convolutional neural networks due to their high dimensionality and complex structures. Current solutions involve performing computations at once, leading to increased latency and decreased energy utilization compared to current approaches like CPU cores. To solve this issue, we propose new techniques called Multiplex Computational Layer Memory Access Machine Learning (MLL) described in detail below.

Images