Home

Installation

Binaries for windows are available as nightly bulds and are for users who would like to try out the software.

If you want to build from source, please follow the instructions to build LichtFeld Studio from source.

Requirements

Software

• OS: Linux (Ubuntu 22.04+) or Windows
• CMake: 3.30 or higher
• Compiler: C++23 compatible (GCC 14+ or Clang 17+)
• CUDA: 12.8 or higher (required)
• LibTorch: 2.7.0 (setup instructions below)
• vcpkg: For dependency management

Hardware

• GPU: NVIDIA GPU with compute capability 7.5+
• VRAM: Minimum 8GB recommended
• Tested GPUs: RTX 4090, RTX A5000, RTX 3090Ti, A100, RTX 2060 SUPER

Build instructions

• Linux
• Windows
• Docker

Usage

The preferred way to use LichtFeld Studio is to import your data (undistorted images + pointcloud + camera locations) in COLMAP format.
Have a look at these 2 introduction videos on how to get your images ready for use in LichtFeld Studio:

Example datasets can be found here

Once your dataset is ready, you can use LFS to train your images to create a Gaussian Splat, either using the GUI or the command line.

• GUI: start LightFeld Studio and use "Import dataset" to load your dataset
• Command line: Basic training:

./build/LichtFeld-Studio -d data/garden -o output/garden

Training with evaluation and visualization:

./build/LichtFeld-Studio \
    -d data/garden \
    -o output/garden \
    --eval \
    --save-eval-images \
    --render-mode RGB_D \
    -i 30000

MCMC strategy with limited Gaussians:

./build/LichtFeld-Studio \
    -d data/garden \
    -o output/garden \
    --strategy mcmc \
    --max-cap 500000

More command line options: command line options