Yunpeng Bai (@Byp215Bai)
2025-10-24 | โค๏ธ 404 | ๐ 50
๐ Do world models need explicit 3D? Our answer: if youโre using Transformers, introducing 3D into DiTโs positional encoding is a natural choice.
๐ Paper: https://arxiv.org/pdf/2510.20385 ๐ HomePage: https://yunpeng1998.github.io/PE-Field-HomePage/ ๐ป Code: https://github.com/MTLab/PE-Field https://x.com/Byp215Bai/status/1981809736535208219/photo/1
๐ Related
Auto-generated bookmark