🔥 Machine learning Docker images for Ruby

Spin up a machine learning environment in a single command

• Docker
• Paperspace
• Binder

Start a container

docker run -p 8888:8888 ankane/ml-stack:standard

And visit https://127.0.0.1:8888/?token=....

Share a volume with the container

-v /path/on/host:/notebooks

For GPU images, install the NVIDIA driver and container toolkit on the host machine. Then use:

--gpus all

Use a custom container for your notebook. Set the container name to the image you want to use, like:

ankane/ml-stack:torch-gpu

And leave the other fields in that section blank.

Create a Dockerfile in the root of your repository with the image you want to use and your notebooks:

# Choose an image
FROM ankane/ml-stack:standard
# Remove example notebooks
RUN rm *.ipynb
# Copy your notebooks
COPY Notebook1.ipynb Notebook2.ipynb ./
# The rest is specific to Binder
ARG NB_USER
ARG NB_UID
RUN adduser --disabled-password --gecos '' --uid ${NB_UID} ${NB_USER} && \
    chown -R ${NB_USER}:${NB_USER} .
USER ${NB_USER}
RUN mkdir ~/.jupyter && \
    echo 'c.KernelSpecManager.ensure_native_kernel = False' > ~/.jupyter/jupyter_notebook_config.py && \
    iruby register

Then visit Binder to run it.

There are a few images to choose from:

• Standard
• Torch GPU

Tag: standard

• Daru
• fastText
• Informers
• LIBMF
• LightGBM
• Numo
• Nyaplot
• Polars
• Prophet
• Rover
• Rumale
• Vega
• XGBoost
• xLearn

Tag: torch-gpu

Designed for deep learning with Torch

• Torch.rb
• TorchAudio
• TorchRec
• TorchText
• TorchVision

Everyone is encouraged to help improve this project. Here are a few ways you can help:

• Report bugs
• Fix bugs and submit pull requests
• Write, clarify, or fix documentation
• Suggest or add new features

To get started with development:

git clone https://github.com/ankane/ml-stack.git
cd ml-stack
# build specific images
docker build -f standard/Dockerfile -t ankane/ml-stack:standard .
docker build -f torch-gpu/Dockerfile -t ankane/ml-stack:torch-gpu .