Welcome to fit-a-nef’s documentation!

fit-a-nef (/fit a nɛf/) is a Python library for quick fitting of thousands of neural fields to entire datasets.

Using the ability of JAX to easily parallelize the operations on a GPU with vmap, a sizeable set of neural fields can be fit to distinct samples at the same time.

The fit-a-nef library is designed to easily allow the user to add their own training task, dataset, and model. It provides a uniform format to store and load large amounts of neural fields in a platform-agnostic way. Whether you use PyTorch, JAX or any other framework, the neural fields can be loaded and used in your project.

Check out the Usage section for further information, including how to Install the library and the dependencies for the repository.

fit-a-nef is developed by the Team at the University of Amsterdam.

Note

Please help us by contributing to the project! See the GitHub repository for more information.

Contents