REAL-TIME DNN EXECUTION FRAMEWORK ON MOBILE DEVICES WITH BLOCK-BASED COLUMN-ROW PRUNING

Number of patents in Portfolio can not be more than 2000

United States of America Patent

APP PUB NO 20230075643A1
SERIAL NO

17797610

Stats

ATTORNEY / AGENT: (SPONSORED)

Importance

Loading Importance Indicators... loading....

Abstract

See full text

BPDNN is a general end-to-end framework to achieve real-time DNN execution on mobile devices. BPDNN supports both CNNs and RNNs. It is based on a novel, fine-grained structured BCR pruning to obtain high execution efficiency without compromising accuracy. BPDNN has two main stages: a compiler-based stage to generate optimized execution codes by leveraging BCR pruning information, and an optimization framework to determine the block size and other hyperparameters based on a decoupling strategy.

Loading the Abstract Image... loading....

First Claim

See full text

Family

Loading Family data... loading....

Patent Owner(s)

Patent OwnerAddress
UNIV NORTHEASTERNNot Provided

International Classification(s)

  • [Classification Symbol]
  • [Patents Count]

Inventor(s)

Inventor Name Address # of filed Patents Total Citations
Li, Zhengang Boston, US 10 16
Niu, Wei Jamestown, US 44 520
Ren, Bin Jamestown, US 114 199
Wang, Yanzhi Newton Highlands, US 10 45

Cited Art Landscape

Load Citation

Patent Citation Ranking

Forward Cite Landscape

Load Citation