Fabrique ton PINNs - Fabrique ton PINNs

This repository is a suite of tutorials to illustrate the PINN’s approach to solving physic problems modeled with Partial Differential Equations with the python programming language.

Pre-requisites¶

Running the tutorials requires an environment with python>=3.10 and the ScimBa library as well as a notebook reader.

We explain below how to meet these requirements with the uv package manager and with pip.

In both cases torch>=2.9 might be downloaded at a point and this requires a good internet connection and some time; we strongly recommend to do it before comming to the workshop.

Note: The first tutorial actually only requires torch>=2.9, matplotlib and numpy, however installing ScimBa will satisfy these dependencies.

In the tutorials we assume the audience to be fluent in python and in numpy.

Meeting the requirements¶

Use either uv (faster) or pip.

Create a directory named fabrique_ton_pinns. In this directory, create and activate a virtual environment:

pip

uv venv
source .venv/bin/activate

Then install the dependencies:

pip

uv pip install scimba jupyterlab

Run the notebooks¶

Start jupyter lab from within the virtual environment:

jupyter-lab

Finally download the desired notebook (.ipynb file) to the fabrique_ton_pinns directory and open it in jupyter lab.