DLC Installation

On IU supercomputer

1. Log onto your favorite computer (e.g., quartz.uit.iu.edu)

   1. Graphically: 

      1. go to https://red.uits.iu.edu

      2. sign in with your IU username and password

      3. Double click on 'terminal' icon 

   2. In a terminal

      1. Open a terminal

      2. ssh <yourIUusername>@quartz.uits.iu.edu

      3. Enter your IU password when prompted for a password

2. Prepare terminal to use Python

[yuma@i51 Quartz]$ module load python/gpu/3.10.0

3. Move to where you want to put the installation package

[yuma@i51 Quartz]$ cd /N/slate/<your IU username>

4. Download a copy of DeepLabCut from github

[yuma@i51 yuma]$ git clone https://github.com/DeepLabCut/DeepLabCut.git

5. Create a Python virtual environment into which you will install DeepLabCut

[yuma@i51 yuma]$ python -m venv /N/slate/<your IU username>/pvenv

6. Activate the virtual environment that you just created

[yuma@i51 yuma]$ source /N/slate/<your IU username>/pvenv/bin/activate

7. Install the Jupyter Notebook package for Python 

(pvenv) [yuma@i51 yuma]$ pip install notebook

8. Install the video editing package ffmpeg

(pvenv) [yuma@i51 yuma]$ pip install ffmpeg

9. Install deeplabcut itself with the gui and tensorflow packages included (** be patient, this step takes ~15 min)

(pvenv) [yuma@i51 yuma]$ pip install deeplabcut[gui,tf]

10. Clean up and leave

(pvenv) [yuma@i51 yuma]$ deactivate

On local computers

Short summary of Installation process:

1. install Anaconda

2. On a conda prompt terminal, call git clone https://github.com/DeepLabCut/DeepLabCut.git

3. Then run: 

   1. cd "DeepLabCut"/

   2. cd conda-environments

   3. conda env create -f DEEPLABCUT.yaml

   4. activate DEEPLABCUT\

For more detailed instruction, look:

1. Has pictures: https://github.com/DeepLabCut/DeepLabCut/tree/master/docs 

2. Has videos: http://www.mackenziemathislab.org/deeplabcut

3. More detailed and in-depth: https://www.nature.com/articles/s41596-019-0176-0

Note: install and setup the latest Anaconda navigator and prompt version; it makes setting up jupyter notebooks, creating virtual environments, and downloading packages easier.

For troubleshooting, look:

1.    https://forum.image.sc/c/usage-issues/7?max_posts=1