This event will comprise a live run-through of a radiology AI model being developed and trained within a container on a SLURM cluster. It will introduce users to the open source imaging toolkit MONAI, cover the basic workflow of developing within containers and show users how containerised applications can be used with common schedulers.

Programme

        Welcome to the event

            Katharina Lauer (ELIXIR)




            Panel Discussion (30-45mins)


            Luciano Guerreiro Lucas (EuroBioimaging Industry board & Leica Microsystems) 

            Rob Holt (Mathworks) 

            Rob Brisk/Nadim Daher (NVIDIA) 

            Bjoern Gruening (ELIXIR) 




            Workshop


                Introduction to Project MONAI (open source medical imaging AI framework)
                Live Demo

                        Introduction to the start environment (Ubuntu with Docker, Nvidia drivers, etc. installed)
                        Pulling a GPU-enabled PyTorch container from NGC (Nvidia’s container repository)
                        Installing additional dependencies inside container (MONAI)
                        Scheduling / running job on a SLURM cluster
                        Writing / running a short application using a Jupyter notebook
                        Q&A





            Rob Brisk/Nadim Daher (NVIDIA)