terminal1/ml repository overview

⁠Machine Learning environment in a box

Machine learning models often require computational libraries for matrix manipulation, optimization or low-level algorithms. Those libraries might require certain system packages (gcc, llvm, openblas, cmake, etc.) and take time to compile. This alpine-based container has a common development environment with a popular python libraries pre-installed.

⁠Quickstart

1. Run a container in a detached mode and share current folder with /home from inside the container:

   host> docker run -it --volume $(pwd):/home --detach --name=ml terminal1/ml
   

   Copy
2. SSH to the container and install anything required. For example, to install tqdm⁠, use the following command:

   host> docker exec -it ml sh
   
   container> pip install tqdm
   

   Copy

⁠Philosophy

• Keep essential individual packages in other reusable containers. This container is unlikely to be a base for anything else.
• Install anything you like to detached container. If installation takes time, use virtualenv⁠ in a shared volume.