Skip to main navigation
Skip to search
Skip to main content
Technion - Israel Institute of Technology Home
Help & Support
Link opens in a new tab
Search content at Technion - Israel Institute of Technology
Home
Researchers
Faculties
Research output
ACCURATE NEURAL TRAINING WITH 4-BIT MATRIX MULTIPLICATIONS AT STANDARD FORMATS
Brian Chmiel
, Ron Banner
, Elad Hoffer
, Hilla Ben Yaacov
,
Daniel Soudry
Electrical and Computer Engineering
Research output
:
Contribution to conference
›
Paper
›
peer-review
Overview
Fingerprint
Fingerprint
Dive into the research topics of 'ACCURATE NEURAL TRAINING WITH 4-BIT MATRIX MULTIPLICATIONS AT STANDARD FORMATS'. Together they form a unique fingerprint.
Sort by
Weight
Alphabetically
Keyphrases
Training Process
100%
Matrix multiplication
100%
Bit Quantization
100%
Bit Matrix
100%
Neural Training
100%
Neural Network Training
50%
Quantized Neural Networks
50%
ImageNet
50%
Deep Neural Network Training
50%
Variance Reduction
50%
Logarithmic Quantization
50%
ResNet50
50%
Quantization Method
50%
Reference Implementation
50%
Engineering
Neural Network Training
100%
Matrix Multiplication
100%
Log Scale
50%
Supplementary Material
50%
Deep Neural Network
50%
Chemical Engineering
Neural Network
100%
Deep Neural Network
100%