The GAN reading list that closed shop in 2017

What it does really-awesome-gan is a curated markdown bibliography of Generative Adversarial Networks resources. It collects papers, tutorials, videos, and reference implementations into categorized lists, from theory and computer vision to what the maintainer calls “Humor.” The whole thing is essentially a single README.md acting as a static index.

The interesting bit The most unusual feature is the project’s explicit retirement. The maintainer stopped adding entries in November 2017 with a frank note: GANs had shifted from an exotic topic to the mainstream, making an exhaustive list “no more feasible or useful.” That makes the repo a fossilized snapshot of the field’s early explosive phase rather than a living document.

Key highlights

• Frozen circa 2017, capturing the foundational GAN era before the literature tsunami.
• Categories span theory, applied vision, applied other domains, and even a “Humor” section.
• Hand-picked recommendations with inline result images (e.g., CycleGAN, face rotation).
• Links out to code repos, including minimalist PyTorch reproductions of classic papers.
• Maintained by Holger Caesar, who also points to related lists like really-awesome-semantic-segmentation.

Caveats

• No new papers have been added since November 2017, so anything recent is absent.
• The maintainer explicitly abandoned the goal of exhaustiveness.

Verdict Worth a look if you are doing historical research or want to trace the GAN family tree from 2014–2017. Skip it if you need a current survey; this is a museum piece, not a roadmap.