# NF2

NF2 is a Python framework for neural non-linear magnetic-field extrapolations in Cartesian and spherical geometries.

## Start Here

New users should begin with the installation guide, run the quickstart example, then choose the extrapolation geometry and training settings for their data.

- [Installation](installation.md) covers package installation and optional dependencies.
- [Quickstart](quickstart.md) walks through the shortest path from input data to an extrapolation.
- [Usage overview](usage.md) shows the command-line, Python, and notebook entry points.

## Main Guides

The main guides collect the common run and analysis pages into a shorter top-level menu.

- [Extrapolations](extrapolations.md) groups Cartesian, spherical, analytical, and series runs.
- [Training](training.md) explains losses, loss schedules, height scaling, batch sizes, validation resolution, and memory handling.
- [Evaluation](evaluation.md) groups export and quality-metric procedures.
- [FAQ](faq.md) answers recurring operational questions.

## Reference Material

Use the reference pages when you need exact configuration keys, command-line options, API signatures, or citation information.

- [Reference](reference.md) groups the API, YAML, dataset, CLI, and export references.
- [Publications](publications.md) lists method and application papers.

```{toctree}
:maxdepth: 2
:caption: Contents

installation
quickstart
usage
extrapolations
downloads
training
evaluation
reference
publications
faq
```