Physics-informed Machine Learning with NVIDIA Modulus
Physics-Informed Machine Learning with NVIDIA Modulus (Latest Version)

Getting started with Modulus Training

This JupyterLab environment has been provided to facilitate your development and experimentation with training Physics-ML models using Modulus. In the left-hand navigation pane, you will see ‘Getting Started with Modulus workflow’ Jupyter Notebook link. You can run the example notebook to understand the recipe for using the Modulus framework for your own use case.

modulus-02.png

Physics-ML model output compared with the analytical result

The training notebook will take you through the key steps:

  • Configuring using Hydra

  • Defining the geometry

  • Specifying the nodes and constraints to set up the computational graph

  • Launching training and validating the results

More detailed description of the example can be found in Modulus documentation here.

© Copyright 2022-2023, NVIDIA. Last updated on Jan 10, 2023.