3D CNN acceleration method and system based on Winograd algorithm
An algorithm and graph technology, applied in the field of 3DCNN acceleration method and system based on Winograd algorithm, can solve problems such as 3DCNN acceleration research
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Example Embodiment
[0056] Example one:
[0057] Such as figure 1 As shown, the implementation steps of the 3D CNN acceleration method based on Winograd algorithm in this embodiment include:
[0058] 1) Read the feature map sub-block Bin to be transformed from the input feature map in, and read the convolution kernel sub-block Bw from the weight buffer w;
[0059] 2) Perform the 3D Winograd algorithm on the feature map sub-block Bin and the convolution kernel sub-block Bw to output the result Tp 1 ;
[0060] 3) Tp for the output result of the 3D Winograd algorithm 1 Carry out accumulation and output the accumulation result Sum;
[0061] 4) Judge whether all the input feature maps in the input feature map in have been read, if it has not been read, skip to step 1); otherwise, skip to step 5);
[0062] 5) Write the accumulation result Sum back to the output feature map buffer Out.
[0063] In this embodiment, when the feature map sub-block Bin to be transformed is read from the input feature map in in step 1)...
Example Embodiment
[0161] Embodiment two:
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap