Abstract: Disclosed herein are a method and apparatus for encoding/decoding a deep neural network. According to the present disclosure, the method for decoding a deep neural network may include: in a plurality of layers of the deep neural network, entropy decoding quantization information for a current layer; performing dequantization on the current layer; and obtaining a plurality of layers of the deep neural network. At least one of global quantization and local quantization is performed on the current layer.
December 11, 2020
January 12, 2023
Korea Electronics Technology Institute, INDUSTRY-UNIVERSITY COOPERATION FOUNDATION KOAEA AEROSPACE UNIVERSITY
Byeong Ho CHOI, Sang Seol LEE, Sung Joon JANG, Sung Jei KIM, Jae Gon KIM, Hyeon Cheol MOON