Shap-E: 生成有条件三维隐式函数（英文版）

英文标题：Shap-E: Generating Conditional 3D Implicit Functions中文摘要：Shap-E 是一种基于条件生成模型的 3D 资产生成方法，其直接生成可以呈现为纹理网格和神经辐射场的隐式函数参数，能够在几秒钟内生成复杂和多样化的 3D 资产，并已通过对 Point-E 模型进行比较，证明其收敛速度更快且显示质量可媲美或更优。英文摘要：We present Shap-E, a conditional generative model for 3D assets. Unlikerecent work on 3D generative models which produce a single outputrepresentation, Shap-E directly generates the parameters of implicit functionsthat can be rendered as both textured meshes and neural radiance fields. Wetrain Shap-E in two s