Everything K-culture — comebacks to K-beauty, straight to your inboxGet it in your inbox

METAL MEDIA

google-gemma/cookbook

3,957Jupyter NotebookApache-2.0

A cookbook repo full of hands-on guides for using Google's Gemma models

google-gemma/cookbook is a GitHub repository collecting tutorials, demo apps, and research notebooks for Google's Gemma family of open models. It covers the core Gemma line as well as specialized variants like CodeGemma for coding, MedGemma for healthcare, and PaliGemma for image understanding. Everything is provided as Jupyter Notebooks, so developers can run the examples directly.

What it does

  1. Covers generations from Gemma, Gemma 2, Gemma 3, Gemma 3n, to Gemma 4, plus purpose-built variants such as CodeGemma, MedGemma, ShieldGemma, and TranslateGemma, each with usage examples.
  2. Organizes content into six folders — Tutorials, Apps (full-stack demos), Experiments (research notebooks), Responsible AI, Docs, and Archive — so readers can find what fits their needs.
  3. Gemma models themselves can be downloaded from Hugging Face Hub, Kaggle, Google Cloud Vertex AI Model Garden, or ai.nvidia.com and used together with these notebooks.
  4. Community members can request new examples via GitHub issues or contribute by picking items off a public wishlist.
  5. The repo has 3,957 GitHub stars, and separate community-maintained translations exist in Traditional and Simplified Chinese.

Why it matters

For developers new to open-weight language models, runnable code examples lower the barrier to entry far more than documentation alone. Since this is Google's own maintained collection, it serves as a trustworthy starting point for anyone building products on top of Gemma.

Terms in this repo

  • Gemma · Google DeepMind's family of lightweight open generative AI models, built on the same research as Gemini
  • Jupyter Notebook · An interactive file format that combines code, explanations, and output in one view
  • CodeGemma · A Gemma variant fine-tuned for coding tasks
  • MedGemma · A Gemma 3-based variant specialized for medical text and image understanding
  • PaliGemma · A vision-language model that jointly understands images and text for deeper image analysis

Repository description (English)

A collection of guides and examples for the Gemma open models from Google.

Owner · google-deepmind

Open on GitHub

Trending repos

All repos →

Latest from METAL MEDIA