notesum.ai

Published at November 26

PassionSR: Post-Training Quantization with Adaptive Scale in One-Step Diffusion based Image Super-Resolution

cs.CV

Released Date: November 26, 2024

Authors: Libo Zhu¹, Jianze Li¹, Haotong Qin², Yulun Zhang¹, Yong Guo³, Xiaokang Yang¹

Aff.: ¹Shanghai Jiao Tong University; ²ETH Zurich; ³Max Planck Institute for Informatics

Arxiv: http://arxiv.org/abs/2411.17106v1

Refer to caption

Components	UNet	VAE	DAPE	ClipEncoder	Total
Params (M)	865,785	83,614	160,335	193,055	1,302,789
MACs (G)	339.241	1,781.123	126.591	14.856	2,261.811