Introduction

DeepLabCut is a powerful Deep Learning-based Python software library for analyzing animal poses from video and image files. It is commonly used in the Biosciences and Neuroscience. Users can install deeplabcut in their home directories using a conda environment from the files provided on the website.

Installing DeepLabCut on your Home Directory

In order to use DeepLabCut on HPC systems, you will need to install it on your home directory. In order to do this, first download the Conda environment YAML file, found Here. Now you need to set up your conda environment. This can be done with the following commands:

module load gnu openmpi cuda anaconda/3.8.3
conda init bash
## Now log out of HPC and log back  in

Your terminal prompt should now look like this: (base) [USER@h22-login-24 ~]$ . From here, you should then create your conda environment using the YAML file you downloaded and then activate it using these steps:

conda env create -f DEEPLABCUT.yaml
## Choose "y" for all prompts
conda activate DEEPLABCUT

You should now be able to start a Python prompt and run this command successfully:

python
## Python will start
## In the python prompt (do not type the >> symbol below)
>> import deeplabcut

Now you should be able to use DeepLabCut! See the GUI Guide and the Usage Paper for more instructions on use of DeepLabCut.

Using the DeepLabCut GUI on HPC Systems

In order to use the GUI for DeepLabCut on HPC's systems, you will need to either login to HPC via ssh using X11 forwarding or you will want to start an RCC Desktop session in Open OnDemand. Information on both of these can be found here:

1. Open OnDemand Documentation
2. SSH Documentation (See the Using Graphical Programs Section)
3. SSH Command: ssh -Y USER@hpc-login.rcc.fsu.edu

Once you have DeepLabCut installed and are logged in using X11 forwarding, run the following commands to get the DeepLabCut Interactive Project Manager GUI running:

conda activate DEEPLABCUT
python -m deeplabcut