1. Install python environment (require: mamba / conda)
Run:

$ mamba env create -p ./.venv -f environment.yml (If running on GPU machine and the appropriate CUDA drivers are installed)

or

$ mamba env create -p ./.venv -f environment-cpu.yml (If running on CPU only machine)

---

2. Source the Python environment:

Run:

$ mamba activate ./.venv

---

3. Run experiments
The experiments can be run on 6 different variants

M1 -> m1
M1' -> m1p
M1'' -> m1pp
M2 -> m2
M2' -> m2p
M2'' -> m2pp

Eg: If you'd like to run experiments for M2', run the following command:

$ ./run.sh m2p

---

4. See results in Jupyter notebook:

Start the jupyter server using:

$ PYTHONPATH=. make jupyter

Open a new browser, navigate to Notebook, change data_root path to the absolute path of ./.output:

Adjust folds, rounds/iterations accordingly in the notebooks.
