@misc{indiciae86a363819030, title = {Training-Free Acceleration of Multi-View 3D Transformers via Regime-Aware Attention Approximation}, author = {Muhammed Izzet Sezer and Sevval Belkis Dikkaya}, year = {2026}, url = {https://searcharxiv.com/abs/2611.17629}, note = {Source identifier: 2611.17629} }