Skip to content

All Pod Variants

Bazzite provides 7 pod variants for different workflows. All are standard OCI containers available at ghcr.io/atrawog/bazzite-ai-pod-*:stable.

Quick Reference

Pod Size GPU Use Case Docker Command
base ~2GB No Web dev, scripting docker run ghcr.io/atrawog/bazzite-ai-pod-base:stable
nvidia ~3GB Yes Custom CUDA setups docker run --gpus all ghcr.io/atrawog/bazzite-ai-pod-nvidia:stable
nvidia-python ~6GB Yes ML/AI with PyTorch docker run --gpus all ghcr.io/atrawog/bazzite-ai-pod-nvidia-python:stable
jupyter ~11GB Yes Interactive notebooks docker run --gpus all -p 8888:8888 ghcr.io/atrawog/bazzite-ai-pod-jupyter:stable
devops ~4GB No AWS, kubectl, Helm docker run ghcr.io/atrawog/bazzite-ai-pod-devops:stable
playwright ~5GB Optional Browser automation docker run -p 5900:5900 ghcr.io/atrawog/bazzite-ai-pod-playwright:stable
githubrunner ~3GB No CI/CD pipelines docker run ghcr.io/atrawog/bazzite-ai-pod-githubrunner:stable

Core Pods

pod-nvidia-python

ML/AI development environment with PyTorch and CUDA support.

  • Size: ~6GB
  • GPU: NVIDIA (CUDA 12.4)
  • Best for: ML model training, deep learning research, PyTorch development
docker run -it --rm --gpus all -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-nvidia-python:stable

Full documentation →

pod-jupyter

JupyterLab server for interactive data science and ML development.

  • Size: ~11GB
  • GPU: NVIDIA (CUDA 12.4)
  • Port: 8888
  • Best for: Data science, interactive ML, notebook-based workflows
docker run -it --rm --gpus all -p 8888:8888 -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-jupyter:stable

Full documentation →

pod-devops

Cloud infrastructure tools - AWS, Google Cloud, Kubernetes, and more.

  • Size: ~4GB
  • GPU: None (CPU-only)
  • Best for: Cloud infrastructure, Kubernetes operations, CI/CD pipelines
docker run -it --rm -v $(pwd):/workspace \
  -v ~/.aws:/home/jovian/.aws:ro \
  -v ~/.kube:/home/jovian/.kube:ro \
  ghcr.io/atrawog/bazzite-ai-pod-devops:stable

Full documentation →

pod-playwright

Browser automation with Playwright, Chrome, and VNC access.

  • Size: ~5GB
  • GPU: Optional
  • Port: 5900 (VNC)
  • Best for: Browser automation, E2E testing, web scraping
docker run -it --rm -p 5900:5900 -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-playwright:stable

Full documentation →

Specialized Pods

pod-base

Foundation pod - Clean Fedora 43 with development essentials.

  • Size: ~2GB
  • GPU: None (CPU-only)
  • Best for: Web development, scripting, testing, Kubernetes work without GPU

Includes:

  • Build toolchain (gcc, make, cmake, ninja)
  • Language runtimes (Python, Node.js, Go, Rust)
  • VS Code, Docker CLI, Podman
  • kubectl, Helm, Claude Code
  • Modern shell tools (fzf, ripgrep, bat, eza)
docker run -it --rm -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-base:stable

pod-nvidia

GPU compute foundation - Adds CUDA toolkit for custom GPU setups.

  • Size: ~3GB
  • GPU: NVIDIA (CUDA 13.0)
  • Best for: Custom GPU compute, CUDA development, building ML frameworks from source

Adds to base:

  • CUDA Toolkit 13.0
  • cuDNN (Deep Neural Network library)
  • TensorRT (inference optimization)
docker run -it --rm --gpus all -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-nvidia:stable

pod-githubrunner

GitHub Actions runner - Self-hosted CI/CD pipeline execution.

  • Size: ~3GB
  • GPU: None (CPU-only)
  • Best for: Self-hosted CI/CD, workflow testing, local GitHub Actions development

Adds to base:

  • GitHub Actions runner agent
  • Kubernetes tools for K8s-based CI
docker run -it --rm -v $(pwd):/workspace \
  ghcr.io/atrawog/bazzite-ai-pod-githubrunner:stable

Pod Selection Guide

Your Need Use This Pod
ML/AI model training nvidia-python
Interactive data science jupyter
Cloud/K8s infrastructure devops
Browser testing playwright
CI/CD pipelines githubrunner
Custom GPU work nvidia
CPU-only development base

See Also