Fourier Neural Operater for Darcy Flow

Core (Latest Release)

This example demonstrates how to set up a data-driven model for a 2D Darcy flow using the Fourier Neural Operator (FNO) architecture inside of Modulus. Training progress can be tracked through MLFlow. This example runs on a single GPU, go to the darcy_nested_fno example for exploring a multi-GPU training.

To train the model, run



training data will be generated on the fly.

Progress can be monitored using MLFlow. Open a new terminal and navigate to the training directory, then run:


mlflow ui -p 2458

View progress in a browser at

If training on a remote machine, set up a ssh tunnel to the server with LocalForward 8080 your_remote_machine_addr:8080. ssh to the server via the specified port, in this case 8080, navigate to the training directory and launch mlflow server


mlflow server --host --port 8080

On your local machine, open a browser and connect to localhost:8080.

Previous MeshGraphNet for transient vortex shedding
Next Nested Fourier Neural Operater for Darcy Flow
© Copyright 2023, NVIDIA Modulus Team. Last updated on Apr 19, 2024.