TurboNeRF founder James Perlman continues to push updates out, this time releasing 0.0.14, with a lot of new features.
Now you can import multiple NeRF datasets
This means you can train multiple NeRFs simultaneously
For now it only runs on a single GPU
Vast improvements in memory management
Duplicating & deleting objects in blender is accurately reflected in the TurboNeRF renderer
Training images can be unloaded to reduce memory usage
Each NeRF now has its own training settings
Ease-of-use improvements in the various UI panels
Reset Training & Unload Training Images has been moved to a new Danger Zone section in the training panel (so you don't accidentally ruin your project!)
Download the TurboNeRF core binary: PyTurboNeRF 0.0.14
Download the Blender addon: TurboNeRF-Blender 0.0.14
James further gave a sneak preview to what's coming next with:
Implementing a faster NeRF scheme (looking at MERF)
Multi-GPU training & rendering
NeRF & Camera effects.
