Release v0.3.0

KhoiDOO released this 05 Jul 13:03

· 875 commits to main since this release

c50a868

Features

Implement differentiable marching cubes CUDA kernel and C++ bindings.
Expose differentiable marching cubes autograd function wrapper and CUDA definitions.
Implement python grid bindings.
Integrate differentiable rendering loss into sparse tensor autoencoder.
Add common differentiable mesh rendering and visualization utilities for notebooks.

Refactors & Build

Refactor and update grid C++ bindings and structures.
Remove old grid.py implementation.
Expose diff marching cubes in ops init and grid structures in data_structure init.
Update standard marching cubes implementation.
Update dependencies and version in pyproject.toml.

Examples

Create diff color marching cubes notebook with voxel color optimization.
Create diff marching cubes notebook with native chamfer distance.
Update chamfer distance, marching cubes, and sparse representation notebooks.
Update README.md.

Assets 2