conda activate ./env

python <file>
python -m pip install <package>
conda install <library>

conda deactivate

Setup

1. Download Miniforge3 (Conda installer) for macOS arm64 chips (M1, M1 Pro, M1 Max).
2. Install Miniforge3 into home directory.

chmod +x ~/Downloads/Miniforge3-MacOSX-arm64.sh
sh ~/Downloads/Miniforge3-MacOSX-arm64.sh
source ~/miniforge3/bin/activate

3. Restart terminal.
4. Make and activate Conda environment. Note: Python 3.8 is the most stable for using the following setup.

conda create --prefix ./env python=3.8
conda activate ./env

5. Install must have packages.

conda install pytorch-nightly::pytorch torchvision torchaudio -c pytorch-nightly
conda install pandas numpy matplotlib scikit-learn notebook

• https://colab.research.google.com/github/ - open any notebook from github in colab select GPU and train faster
• jupyter nbconvert --clear-output --inplace notebook.ipynb - clear notebook output in all cells