ICGui   

Graphical User Interface for the NeRFICG framework

---

Getting Started

Ensure NeRFICG has been installed and the appropriate conda environment is activated.

Example Usage

Standalone Viewer

The standalone viewer renders a pre-trained model from a checkpoint file. By running the command below, a launcher is displayed to choose the model directory (or training config and checkpoint separately) and configure the GUI.

python ./scripts/gui.py  # from the base directory of the NeRFICG repository

The model to render as well as any GUI settings can also be passed on the command line; for more info run

python ./scripts/gui.py --help

Training Viewer

The model can also be viewed during training. This is done by simply enabling the GUI in the config file before launching scripts/train.py. The GUI will start automatically when training begins. Alternatively, you can also enable the GUI using

python ./scripts/train.py -c ./configs/config.yaml TRAINING.GUI.ACTIVATE=True

Depending on the frequency of rendering set in the config, this may slow down training! The GUI can be closed at any point during training without affecting the training process to free up resources used by the GUI.

Acknowledgments

Special thanks go out to the following projects:

• torchwindow, for demonstrating how to directly render tensors without copying them to the CPU.