For this project we will be using the Anaconda Package Manager for python. If you haven't used Anaconda before, this is a great time to learn!
To install anaconda on the server you'll need to download and run the installer:
cd /tmp
wget https://repo.anaconda.com/archive/Anaconda3-2020.02-Linux-x86_64.sh
bash Anaconda3-2020.02-Linux-x86_64.shAfter anaconda is installed you can install the project dependencies by creating a new covid environment.
If you cloned this repo to your home directory this would look like this:
cd ~/covid-blood
conda env create -f environment.ymlYou can check if it worked by a) looking at the logs, and b) by running conda activate covid
We will be using wandb to track experiments. You can sign up for a free account here: Weights and Biases. You'll want to use your .edu email to get all the free stuff :).
When you first run an experiment you'll be prompted to authorize your account and be told where to look to track your experiments.
Due to IRB guidelines all our data will remain on Duke controlled computers. For now that will mean the COL GPU Server.
To work with this data you will need to remotely access the server either through SSH or RDP.
PyCharm has a great remote debugging feature built into it's professional edition. Remote debugging allows you to run your code on a remote server, but debug it using your local IDE. You can find instructions for setting this up here: Remote Debugging over SSH in PyCharm
Don't push to master!
Functions should use named arguments by default