Seventeen thousand stars, one giant list

What it does This repository is an awesome-list: a manually curated directory of external TensorFlow resources. It links out to tutorials, model implementations, libraries, utilities, videos, papers, and books, each tagged with a one-line description. Think of it as a human-powered search engine for the graph-framework era.

The interesting bit The entire product is editorial judgment. Someone waded through hundreds of GitHub repos, Coursera courses, and arXiv pre-prints so you can find a Raspberry Pi build, a WaveNet clone, or a Shakespeare translator without opening fifty tabs first.

Key highlights

• Coverage is encyclopedic: beginner notebooks, Stanford’s CS20SI syllabus, Android deployment guides, reinforcement-learning agents, and 3D convolutional nets all share the same page.
• Heavy on reproducible research: links include YOLO, DCGAN, Neural Turing Machines, and attention-based captioning, most paired with their original paper citations.
• Surprises in the long tail: you’ll find everything from Single Image Random Dot Stereograms to DeepOSM (OpenStreetMap satellite training) and monolingual Shakespeare-to-modern-English translation.
• Organized by category—tutorials, models, libraries, tools—rather than by difficulty or date, which rewards browsing but assumes you know what you want.

Caveats

• The README is long enough that GitHub truncates the raw view; you’ll need in-page search or the table of contents to navigate.
• Link descriptions range from detailed paragraphs to bare repo names, so quality and context vary line by line.
• Many entries reference specific 2015–2017 academic implementations; treat it as a historical archive as much as a current index.

Verdict Worth a scroll if you are hunting for a specific reference implementation of a classic paper or need a curated jump-off point into the TensorFlow ecosystem. If you expect actively maintained, copy-paste-ready code for modern APIs, look elsewhere—or at least check the linked repo dates first.