Skip to content

datascience images

Lightweight data science environment for intro Python and ML courses. Three variants: NVIDIA GPU (with CuPy acceleration), CPU-only, and Apple Silicon.


datascience-nvidia

Component Version
Base image nvidia/cuda:12.8.1-runtime-ubuntu24.04
Python 3.12
CUDA 12.8
GPU support Pascal-Blackwell (sm_60-sm_100)

Packages: numpy, pandas, scipy, scikit-learn, xgboost, statsmodels, matplotlib, seaborn, plotly, jupyterlab, optuna, python-dotenv, cupy 13.6.0 (custom wheel)

DockerHub: gperdrizet/datascience-nvidia


datascience-cpu

Component Version
Base image python:3.12-slim
Python 3.12

Packages: numpy, pandas, scipy, scikit-learn, xgboost, statsmodels, matplotlib, seaborn, plotly, jupyterlab, optuna, python-dotenv

DockerHub: gperdrizet/datascience-cpu


datascience-mac

Native linux/arm64 image for Apple Silicon (M1/M2/M3). Docker Desktop on Mac runs it without emulation. GPU/Metal passthrough is not supported inside Docker on macOS.

Component Version
Base image python:3.12-slim
Platform linux/arm64
Python 3.12

Packages: numpy, pandas, scipy, scikit-learn, xgboost, statsmodels, matplotlib, seaborn, plotly, jupyterlab, optuna, python-dotenv

DockerHub: gperdrizet/datascience-mac