Shap-E: Generating Conditional 3D Implicit Functions

ABSTRACT

We present Shap-E, a conditional generative model for 3D assets. Unlike recent work on 3D generative models which produce a single output representation, Shap-E directly generates the parameters of implicit functions that can be rendered as both textured meshes and Neural Radiance Fields (NeRFs). We train Shap-E in two stages: first, we train an encoder that maps 3D assets into the parameters of an implicit function; second, we train a conditional diffusion model on these continuous latent representations. Shap-E converges faster and reaches comparable or better sample quality than Point-E, while natively supporting high-quality rendering and mesh extraction from text prompts in a matter of seconds.