If you prefer developing jupyter notebooks with Visual Studio Code instead of the native Jupyter Notebooks interface, this is for you. It runs a python3 container (with an ipython kernel installed) with the most common libraries like pandas, matplotlib, scikit-learn or tensorflow.
Simply have the Dev Containers installed in Visual Studio Code and open the repository. Everything else will be prompted VS Code.
If you plan to leverage your GPU for training, make sure you have the NVIDIA Container Toolkit installed and you remove the comment in front of "runArgs": ["--gpus", "all"]
.