How to run a use case

First, create the use case’s Python environment (i.e., PyTorch or TensorFlow) as described here, and activate it. Then, install use case-specific dependencies, if any:

pip install -r /use/case/path/requirements.txt

Alternatively, you can use the use case Docker image, if available.

Then, go to the use case’s directory:

cd /use/case/path

From here you can run the use case (having activated the correct Python env):

# Locally
python train.py [OPTIONS...]

# With SLURM: stdout and stderr will be saved to job.out and job.err files
sbatch startscript

Fast particle detector simulation | CERN use case

The first interTwin use case integrated with itwinai framework is the DT for fast particle detector simulation. 3D Generative Adversarial Network (3DGAN) for generation of images of calorimeter depositions. This project is based on the prototype 3DGAN model developed at CERN and is implemented on PyTorch Lightning framework.

• 3DGAN
  • itwinai x 3DGAN
  • interLink x 3DGAN

MNIST dataset use case

MNIST image classification is used to provide an example on how to define an end-to-end digital twin workflow with the itwinai software.

• MNIST
  • Torch Lightning
  • PyTorch
  • Tensorflow

Tropical Cyclones Detection | CMCC use case

Below you can find the training and validation of a Tropical Cyclones (TCs) Detection model, developed by CMCC, integrated with itwinai framework.

• Tropical Cyclones Detection

Noise Simulation for Gravitational Waves Detector (Virgo) | INFN use case

Below you can find the integration of the Virgo use case with itwinai framework, developed by INFN.

• Virgo