Modules in Holmes Lab Conda

Last updated: May 19, 2025 11:01 AM

Activate Holmes Lab Conda

In the terminal, signed into Amarel, run:

conda activate /projects/community/holmesenv

Now you are ‘in’ the Holmes Lab conda environment. Anything you install using pip install ... or conda install ... will be installed to this conda.

To de-activate Holmes Lab Conda, run:

conda deactivate

To use Holmes Lab Conda Packages (or another env) in Slurm Scripts:

Go to your terminal, and run:

$ nano ~/.bashrc

Add the following lines to your bashrc– this will mean that every time you sign into Amarel, your account will automatically activate the Holmes Lab Conda. Thus, any slurm scripts will also use the Holmes Lab Conda

# Set up for conda
. /projects/community/py-data-science-stack/5.1.0/kp807/etc/profile.d/conda.sh

# Activate conda
source /projects/community/py-machine-learning/intel18/cuda12/pgarias/etc/profi$
conda activate /projects/community/holmesenv

If you want to automatically activate a different conda, just replace /projects/community/holmesenv with your conda environment name.

In any slurm scripts, add this before your command, to make sure your .bashrc is loaded:

source ~/.bashrc

Modules in Holmes Lab Conda

To see all the current modules in our conda, type:

# While already actiated /projects/community/holmesenv
conda list

Add to Holmes Lab Conda

To install packages or modules, just run:

pip install packageName
# or
conda install packageName

Conda install is safer, but only big / well known packages are available in conda install, so pip is also fine.

Create your Own Conda based off Holmes Lab Conda

In order to duplicate this environment excute the following:

  1. conda activate MyENV (where MyENV is the environment name)

  2. conda list --explicit > spec-file.txt This will produce spec-file.txt that can be used to replicate the environment like this:

  3. conda create --name NewEnv --file spec-file.txt
  4. conda activate NewEnv

Now you are inside of your personal conda environment, which is a duplicate of the Holmes Lab environment. If you add packages here, they won’t be added to the

What is a Conda Environment?

A Conda environment is an isolated workspace that contains a specific set of software packages, libraries, and a Python version — all configured to work together. It allows researchers to run code without worrying about software conflicts or setup differences between computers.

How We Use Conda in the Holmes Lab

To keep things simple and consistent across the team, we maintain a shared Conda environment that includes all the core packages commonly used in our research (e.g., for data analysis, neuroimaging, statistics, etc.).

Instead of each person manually installing packages on their own system, everyone can activate the Holmes Lab Conda environment and start working immediately with the same setup. This has several key benefits:

  • No individual setup required – saves time and reduces setup errors.
  • Consistency across users – ensures code runs the same way on everyone’s machine or the HPC cluster.
  • Centralized maintenance – when we update or add packages, it only needs to be done once.
  • Reproducibility – shared environments make it easier to reproduce results and share notebooks/scripts.