The map of the PyTorch ecosystem, one giant link list at a time

What it does It’s a link directory, not code: a hand-curated index of PyTorch-adjacent GitHub projects organized by domain — NLP and speech, computer vision, probabilistic/generative libraries, plus tutorials, books, paper implementations, and conference talks. Think of it as a card catalog for the ecosystem rather than a framework.

The interesting bit Awesome-lists are glue by definition, but this one earned ~16.7k stars by being the de facto index before PyTorch had an official ecosystem page. It’s also a time capsule: entries like BERT-PyTorch and maskrcnn-benchmark sit alongside still-current staples like transformers and kornia.

Key highlights

• Domain-sorted sections: NLP/speech, CV, probabilistic/generative, and general libraries
• Dedicated sections for paper implementations and tutorials/books
• Covers talks and conferences, plus “PyTorch elsewhere” for non-GitHub resources
• Contributions are explicitly welcome, per the README badges

Caveats

• Curation ≠ maintenance: some linked projects are archived or long dead, and the list itself shows signs of age (its own star badge still says 12,400+)
• No annotations beyond one-line descriptions — you still have to vet each link yourself

Verdict Useful as a starting point if you’re surveying the PyTorch landscape or hunting for a paper implementation. Skip it if you already know exactly which library you need — just go there directly.