DeepLabCut

GUI

The DeepLabCut GUI can be accessed via the Open OnDemand Applications. It is highly recommended to run this on a GPU compute/CUDA partition.

Command line interface

DeepLabCut is made available on the cluster as a shared Apptainer container image. Scripted/commandline access requires Python to be run within the context of the container.

• Open a terminal on one of the GPU nodes.

• Start ipython in the container

apptainer run --nv /opt/apptainer_img/deeplabcut-2.3.9-cuda11.8.sif ipython

• Then type: import deeplabcut

• See https://deeplabcut.github.io/DeepLabCut/docs/standardDeepLabCut_UserGuide.html#deeplabcut-in-the-terminal-command-line-interface

Slurm

DeepLabCut is made available on the cluster as a shared Apptainer container image. Scripted/commandline access requires Python to be run within the context of the container.