PyTorch Summer Hackathon Winners
summary
Menlo Park hosts a 28‑hour hack with 74 participants; the global event runs five weeks and attracts about 1,500 submissions. First‑place build is a usable meta‑learning library learn2learn that implements multiple state‑of‑the‑art algorithms, test suites and PyTorch‑compatible APIs. Runners‑up include an exoplanet discovery pipeline, an end‑to‑end music‑mixing system using torchaudio, a StyleGAN T‑shirt generator, and a block‑based deep‑net kit for kids — all shipped and demoed successfully.
AI-written summary · about
chapters (5)
- 0:00PyTorch Hackathon intro
- 1:04Hackathon winners projects
- 2:54Meta-learning library: Learn to Learn
- 7:56Torch Meta for research
- 10:46Closing remarks
related talks
22mHackathon Winners Presenting On Stage
02mPyTorch Conference 2025 Recap
58mImproving disaster response with machine learning
01h 32mIntroduction to PyTorch
46mHackathon Presentations
05mTips and Tricks on Hacking with PyTorch: A Quick Tutorial
12mLightning Talk: State of PyTorch
43mPyTorch Technical Deep Dive