1 paper
Xiao Luo, Mingyang Du, Xin Zhou +5
High-fidelity 3D generation predominantly relies on scaling model capacity and data, which incurs prohibitive computational costs. This paradigm typically requires learning geometr…