ADFQ-ViT: Activation-Distribution-Friendly Post-Training Quantization for Vision Transformers.

Yanfeng Jiang Ning Sun Xueshuo Xie Fei Yang Tao Li

Published in: CoRR (2024)

Keyphrases

computer vision
vision system
training process
real time
training and test data
training set
training algorithm
probability distribution
training samples
quantization error
data distribution
training examples
image processing
gaussian distribution
multiresolution
spatial distribution
data sets