Release v0.4.5

Features:

• Bind PyTorch tensor methods for newly added mesh geometry metrics via pybind11
• Implement parallel CUDA kernels and host wrappers for geometric quality metrics
• Declare mesh-level methods for extended geometric quality metrics in TriangleMesh
• Add robust primitive methods for geometry metrics computation in Triangle struct
• Add FID script for generation evaluation

Refactors & Build:

• Bump version to 0.4.5
• Update transformer architecture for point cloud generation
• Update point cloud training script
• Update point cloud generation script
• Update classification models
• Update citations and references for surface remeshing and marching cubes
• Add various .json tracking files (fid.json, history.json)
• Update Dockerfile and docker-compose.yml
• Update git attributes

Examples:

• Add main script for marching tetrahedra