Evict3R: Training-Free Token Eviction for Memory-Bounded Streaming Visual Geometry Transformers

作者：Soroush Mahdi, Fardin Ayar, Ehsan Javanmardi, Manabu Tsukada, Mahdi Javanmardi · 单位：Amirkabir University of Technology (AUT), Tehran, Iran, The University of Tokyo, Tokyo, Japan · 会议/期刊：arXiv preprint · 方向：6D位姿 · 发布日期：2025-09-22