AI/ML Docker Images

Overview

This repository contains a collection of AI/ML Docker images catering to different levels of development and deployment needs. These images are preconfigured with essential tools, libraries, and dependencies to streamline machine learning workflows.

Available Images

1. Minimal AI/ML Tools

• Image: minimal-aiml
• Description: A lightweight image with essential AI/ML tools for data processing and scientific computing.
• Base Image: python:3.9
• Tags: latest, minimal
• Includes:
  • Python Essentials: pip, setuptools
  • Numerical & Data Processing: numpy, pandas
  • Visualization: matplotlib, seaborn
  • Jupyter Support: jupyter
  • Utilities: scipy, tqdm
• Build and Push:
  • Navigate to the minimal directory and run the following command:

    ./build_and_push.sh

2. Common AI/ML Tools

• Image: common-aiml
• Description: Includes TensorFlow, PyTorch, scikit-learn, and Jupyter Notebook.
• Base Image: python:3.9
• Tags: latest, common
• Includes:
  • Python Essentials: pip, setuptools, wheel
  • Numerical & Data Processing: numpy, pandas, scipy, tqdm, joblib
  • Visualization: matplotlib, seaborn, plotly
  • Machine Learning: scikit-learn, xgboost, lightgbm, catboost
  • Deep Learning Frameworks: tensorflow, torch, torchvision, torchaudio, keras
  • Jupyter & Interactive Development: jupyter, jupyterlab, notebook, ipython, nbconvert
  • Data Handling & Feature Engineering: pillow, opencv-python, nltk, spacy, transformers
  • Model Persistence & Deployment: mlflow, onnx, onnxruntime, fastapi
• Build and Push:
  • Navigate to the common directory and run the following command:

    ./build_and_push.sh

3. Advanced AI/ML Tools

• Image: advanced-aiml
• Description: Contains AI/ML frameworks with GPU acceleration, Hugging Face Transformers, XGBoost, and MLFlow.
• Base Image: nvidia/cuda
• Tags: latest, advanced

4. Kitchen Sink (Full AI/ML Stack)

• Image: kitchen-sink-aiml
• Description: A comprehensive AI/ML environment with all major frameworks, libraries, and Jupyter support.
• Base Image: nvidia/cuda
• Tags: latest, full-stack

Getting Started

Pull an Image

# Example: Pull Common AI/ML Tools Image
docker pull dockerhub.com/capturealpha/common-aiml:latest

Run a Container

# Example: Run Jupyter Notebook with Advanced AI/ML Tools
docker run --gpus all -p 8888:8888 -v $(pwd):/workspace dockerhub.com/capturealpha/advanced-aiml

Build Your Own Image

Each directory contains a Dockerfile. To build an image locally:

cd common-aiml
docker build -t my-common-aiml:latest .

Contributing

Contributions are welcome! Feel free to submit a PR for new AI/ML Docker images or improvements to existing ones.

License

This repository is licensed under the MIT License. See LICENSE for details.

Maintainers

• Nathan Bolam - GitHub