o76923/xgboost repository overview

⁠XGBoost Stack

This image creates a stack for use with nvidia-docker that builds on the CUDA stack and adds Intel's Python 3 distribution as well as numpy, pandas, joblib, and scikit-learn in order to have a robust base for running XGBoost efficiently with both MKL optimized math operations and the ability to use a GPU for running XGBoost itself.

⁠Using this image

In order for this image to work properly, you must use nvidia-docker instead of just using docker.

When using this as a base image, the default entrypoint is an executable located at /app/sh/entrypoint.sh.