Skip to content

Latest commit

 

History

History
9 lines (5 loc) · 507 Bytes

Readme.md

File metadata and controls

9 lines (5 loc) · 507 Bytes

Reconstructing 3D Shapes from the ShapeNet Dataset using a Patch Vector Quantized Variational Autoencoder

This Repo is a PyTorch implementation of the PVQVAE from the paper "3DQD: Generalized Deep 3D Shape Prior via Part-Discretized Diffusion Process" by Wang et al.

Paper

Major differences and deviations from the paper are discussed in the poster below.

3D Reconstruction poster