English(EN) QuantSR+: Pushing the Limit of Quantized Image Super-Resolution Networks

QuantSR+ 框架提升低比特量化图像超分辨率性能

作者 PulseAugur 编辑部 · [2 个来源] · 2026-05-21 11:38

研究人员开发了 QuantSR+，一个旨在提升使用低比特量化的图像超分辨率模型性能的新框架。该方法解决了模型压缩到 2-4 比特时常出现的显著性能下降问题。QuantSR+ 在量化算子、网络架构和训练策略方面进行了改进，以在准确性和效率之间取得更好的平衡。 AI

影响提高了压缩图像超分辨率模型的效率和准确性，使其能够在资源受限的设备上更广泛地部署。

排序理由该集群包含一篇详细介绍图像超分辨率新技术的学术论文。

AI 生成摘要 · Google Gemini · 来自 2 个来源。我们如何撰写摘要 →

报道来源 [2]

arXiv cs.CV TIER_1 English(EN) · Haotong Qin, Xudong Ma, Xianglong Liu, Jie Luo, Jinyang Guo, Michele Magno, Yulun Zhang · 2026-05-22 04:00

QuantSR+: 突破量化图像超分辨率网络的极限

arXiv:2605.22351v1 Announce Type: new Abstract: Low-bit quantization is widely used to compress super-resolution (SR) models and reduce storage and computation costs for deployment on resource-limited devices. However, when SR models are pushed to ultra-low precision (2-4 bits), …
arXiv cs.CV TIER_1 English(EN) · Yulun Zhang · 2026-05-21 11:38

QuantSR+: 突破量化图像超分辨率网络的极限

Low-bit quantization is widely used to compress super-resolution (SR) models and reduce storage and computation costs for deployment on resource-limited devices. However, when SR models are pushed to ultra-low precision (2-4 bits), performance can drop sharply due to diminished r…