H2O's official tutorial dump: 1.5k stars, zero hand-holding

What it does This repo collects training materials for H2O-3, the open-source ML platform. You’ll find Jupyter notebooks and R scripts covering introductions, grid search, deep learning, stacked ensembles, and AutoML. The master branch tracks the latest stable H2O release; named branches freeze versions for past H2O World conferences (2015, 2017).

The interesting bit The historical snapshots are the unusual angle — if you’re stuck on an old H2O release at work, the Tibshirani-3 or Wheeler-2 branches may save you from version-mismatch hell. That’s more foresight than most tutorial repos bother with.

Key highlights

• Dual-language coverage: Python notebooks and R scripts for most core topics
• AutoML and stacked ensembles tutorials included (the newer H2O features)
• Event-branched history: 2015 and 2017 H2O World materials pinned to their release
• requirements.txt provided for Python dependency setup
• Points to Stack Overflow and Google Groups for support rather than managing issues here

Caveats

• README is a link index, not a curriculum — you’ll need to hunt through SUMMARY.md for structure
• Some deep-learning material lives in external repos (ledell’s sldm4-h2o, LatinR 2019)
• No visible CI or testing on the notebooks; “should work” is the guarantee

Verdict Worth bookmarking if you’re already committed to H2O-3 and want vendor-authored examples. Skip it if you need a guided course or are evaluating frameworks — this is reference material, not persuasion.