EFloat: Entropy-coded Floating Point Format for Deep Learning.

Rajesh Bordawekar Bülent Abali Ming-Hung Chen

Published in: CoRR (2021)

Keyphrases

floating point
deep learning
unsupervised learning
lattice vector quantization
fixed point
subband coding
machine learning
instruction set
three dimensional
floating point arithmetic
mental models
vector quantizer
weakly supervised
object recognition
filter bank
coding scheme