Sign in
A 4-bit Integer-Only Neural Network Quantization Method Based on Shift Batch Normalization.
Qingyu Guo
Xiaoxin Cui
Jian Zhang
Aifei Zhang
Xinjie Guo
Yuan Wang
Published in:
ISCAS (2022)
Keyphrases
</>
neural network
preprocessing
high precision
computational complexity
computational cost
high accuracy
significant improvement
bp neural network
dynamic programming
fuzzy neural network
training process
prediction model
support vector machine
cost function
artificial neural networks
pairwise
genetic algorithm