16 hrs
On-demand video
PyTorch-only
Simple dependencies
No CUDA
Research friendly code
What You'll Learn
- Parse COLMAP outputs for camera poses and sparse reconstructions
- Implement 3D Gaussian primitives as scene representations
- Build a trainable neural rendering pipeline modeling view-dependent radiance
- Code the full training and optimization loop for 3D Gaussian Splatting
- Create a differentiable renderer producing photorealistic images
- Explore visualization techniques for complex 3D data and 3DGS models
Who This Course Is For
- Researchers and grad students wanting full algorithmic understanding of 3DGS
- Developers seeking clean, flexible PyTorch implementations
- Anyone avoiding C++/CUDA complexity while achieving SOTA performance
- Those building foundations in neural rendering and 3D computer vision
Prerequisites
- Python programming experience
- Basic familiarity with PyTorch and NumPy
What Students Say About My NeRF Course
I’ve had the privilege of teaching NeRF to over 800 students worldwide. My NeRF course is rated ⭐ 4.5 / 5 based on 150+ verified reviews.
