L’interface est localisée ; les présentations et documents des Skills sont traduits automatiquement par IA dans votre langue. Le code, les commandes, chemins et termes techniques restent en version originale, et le document de l’auteur fait foi.
Explorer les Skills
Productivité · SKILL.md

viewers

viewers

Guides gaussian-splatting SIBR remote and real-time viewer build, run, connection, and troubleshooting workflows.

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentation de la Skill

GitHub

name: viewers description: "Guides gaussian-splatting SIBR remote and real-time viewer build, run, connection, and troubleshooting workflows." disable-model-invocation: true metadata: disco-role: operating license: NOASSERTION

Viewers

Use this sub-skill when the task is about the SIBR network viewer, the real-time Gaussian viewer, viewer build commands, optimizer connection settings, navigation, top view, or CUDA/OpenGL interop issues.

Read First

Viewer Types

  • Network/remote viewer connects to a running train.py optimizer over a socket. The optimizer defaults to 127.0.0.1:6009 unless --ip and --port are changed.
  • Real-time viewer loads a trained model directory and renders it interactively using SIBR/OpenGL/CUDA.

Preconditions

  • Python training/rendering setup is handled by ../setup-and-backends/SKILL.md.
  • A trained model for the real-time viewer is produced through ../training/SKILL.md.
  • SIBR viewer binaries must already be downloaded or built; this sub-skill does not assume they exist.

Command Shapes

Remote viewer:

SIBR_remoteGaussian_app --ip 127.0.0.1 --port 6009

Optimizer side:

python train.py -s <scene> --ip 127.0.0.1 --port 6009

Real-time viewer:

SIBR_gaussianViewer_app -m <trained-model>

Use -s <source-scene> if the viewer cannot resolve the model's saved source path.

Boundaries

  • Do not diagnose SIBR internals beyond documented build/run and runtime failure modes unless the user is explicitly editing the SIBR source.
  • Do not treat SIBR viewer success as proof that Python training/rendering metrics work, or vice versa.
Skills enregistrées