NeRF-playground

This is a pytorch-based repository with implementation of the Neural Radiance Field (NeRF). All implementations have a detailed description.

Current implementations:

• Vanilla NeRF
• Enhanced version of Vanilla NeRF (tiny-cuda-nn)

---

Requirements

To run the code from this repository the following is essential:

• To run Vanilla NeRF training NVIDIA GPU will be preferable with 8 GB of RAM minimum. However, the code can be run on CPU too.
• Conda environment: needs to be installed and initialized using environment.yml file. All dependencies will be automatically installed.
• install tiny-cuda-nn using the following command:

pip install git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch

---

Project description

The repository contains two tools:

• TrainNerf_tool.py - script for training vanilla NeRF model
• MeshNerf_tool.py - script for extracting mesh from the NeRF model. This is a very simple meshing algorithm based on mcubes package.

To run both scripts you will have to open them in editor and tune parameters according to your needs.