AISetup
Overview
These scripts will automatically set up local python environment for `PyTorch` and `JAX` for deep learning.
Instructions:
- Download the folder at the bottom of this page.
- Copy it to your `$WORK` directory on the FAU cluster:
scp -r AISetup alex:/home/woody/ihpc/YOUR_FAU_USERNAME/
- Login to `alex`:
ssh alex
- Navigate to your $WORK directory:
cd $WORK
- Unzip the contents:
unzip AISetup-20250522.zip -d nhr_grad_school_2025
- Navigate to the setup folder:
cd nhr_grad_school_2025
- Source the Miniforge installation script:
source prepare_python_env.sh
- Source the PyTorch environment setup script:
source prepare_pytorch_env.sh
- Run the PyTorch dataset script:
mamba activate pytorch; python prepare_pytorch_data.py
- Source the JAX environment setup script:
source prepare_jax_env.sh
- Run the JAX dataset script:
mamba activate jax; python prepare_jax_data.py
This will setup up a local Python installation in your $WORK directory, create independent environments for PyTorch and JAX installations, and download/prepare necessary datasets. After running all the preparation scripts you can check that things are set up properly:
(jax) ihpc153h@alex2:/home/woody/ihpc/YOUR_FAU_USERNAME$ mamba info --envs
Name Active Path
──────────────────────────────────────────────────────────────────────
base /home/woody/ihpc/ihpc153h/miniforge3
jax * /home/woody/ihpc/ihpc153h/miniforge3/envs/jax
pytorch /home/woody/ihpc/ihpc153h/miniforge3/envs/pytorch
To switch between environments, simply run:
mamba activate YOUR_DESIRED_ENVIRONMENT
⚠️🚨 IMPORTANT🚨 ⚠️
Initializing `mamba/conda` adds the following lines to your `$HOME/.bashrc`:
export MAMBA_EXE='/home/woody/ihpc/YOUR_FAU_USERNAME/nhr_grad_school_2025/miniforge3/bin/mamba';
export MAMBA_ROOT_PREFIX='/home/woody/ihpc/YOUR_FAU_USERNAME/miniforge3';
__mamba_setup="$("$MAMBA_EXE" shell hook --shell bash --root-prefix "$MAMBA_ROOT_PREFIX" 2> /dev/null)"
if [ $? -eq 0 ]; then
eval "$__mamba_setup"
else
alias mamba="$MAMBA_EXE" # Fallback on help from mamba activate
fi
unset __mamba_setup
# <<< mamba initialize <<<
This sets your Python installation to the one that we have created everytime you open a new shell on the FAU cluster. We recommend that you remove these lines after the course concludes to avoid future issues with Python modules provided by FAU.
This can be done easily by running `mamba_clean.sh`:
bash mamba_clean.sh