Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

A method and apparatus for unwrapping tensor data of a convolution neural network

A convolutional neural network and data technology, applied in the field of convolutional neural network, can solve the problems that cannot be processed, convolutional neural network accelerator cannot be processed directly, and cannot be processed efficiently, so as to achieve the effect of improving processing efficiency

Active Publication Date: 2019-01-11
HORIZON ROBOTICS SHANGHAI ARTIFICIAL INTELLIGENCE TECH CO LTD
View PDF14 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, the convolution operations actually performed by the convolutional neural network accelerator may be various, for example, the input tensor data and weight parameters may have various shapes, and the convolution operations to be performed may have various shapes. kind of step size
The diversity of data and operations prevents convolutional neural network accelerators from processing directly or efficiently, or even at all

Method used

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
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A method and apparatus for unwrapping tensor data of a convolution neural network
  • A method and apparatus for unwrapping tensor data of a convolution neural network
  • A method and apparatus for unwrapping tensor data of a convolution neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0019] As mentioned above, due to many considerations and constraints such as chip space, hardware cost, design complexity, etc., convolutional neural network accelerators are usually designed to be able to process shapes that conform to certain specifications (for example, with specified width, height and / or number of channels) and / or canonical quantities of tensor data and convolution kernels, and convolution operations with certain canonical strides are supported.

[0020] In order to cope with the diversity of data and operations, additional logic control circuits can be set in the convolutional neural network accelerator, for example, to control the selection and timing control of multipliers and adders. However, this increases the design cost and complexity of the hardware, and may cause waste of hardware resources due to the idleness of the multiplier-adder. Moreover, even if such settings are added to the convolutional neural network accelerator, it may still be imposs...

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

PUM

No PUM Login to View More

Abstract

A method and apparatus for unwrapping tensor data of a convolution neural network are disclosed. The method comprises the following steps: determining an expansion dimension according to the first tensor and the hardware specification of the convolution neural network accelerator; determining a second tensor based on the first tensor; and determining a second convolution kernel for the second tensor from the first convolution kernel for the first tensor. By the method and apparatus according to the embodiments of the present disclosure, data and operation parameters in a canonical form conforming to the requirements of a convolution neural network accelerator can be obtained, thereby greatly improving the processing efficiency of the convolution neural network accelerator.

Description

technical field [0001] The present disclosure generally relates to the technical field of convolutional neural networks, and in particular to methods and apparatus for unrolling tensor data of convolutional neural networks. Background technique [0002] Deep learning technology based on convolutional neural network has been widely used in different fields such as image recognition, video analysis, natural language processing, and assisted driving. The computational load in convolutional neural networks is usually very large. It is desirable to be able to efficiently perform operations in convolutional neural networks using hardware such as general-purpose central processing units (CPUs), graphics processing units (GPUs), or dedicated accelerators. [0003] Due to many considerations and constraints such as chip space, hardware cost, and design complexity, convolutional neural network accelerators (e.g., may include arrays of multiply-add cells) are usually designed to be ab...

Claims

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

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06N3/063
CPCG06N3/063
Inventor 李德林凌坤李建军王振江
Owner HORIZON ROBOTICS SHANGHAI ARTIFICIAL INTELLIGENCE TECH CO LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com