ayaanzhaque/instruct-nerf2nerf
A neural radiance field editing system that uses diffusion models to modify 3D scenes based on text instructions.

Velocity · 7d
+0.7
★ / day
Trend
→steady
star history
Instruct-NeRF2NeRF is an ICCV 2023 paper implementation that iteratively updates 3D scene representations using text-guided diffusion models. Built on Nerfstudio and PyTorch, it allows users to edit trained NeRF scenes with natural language commands by projecting image-level edits from a conditioned diffusion model back into the 3D volume.