Halcyon

Search everything

Mentions · 7d0
Mentions · 30d1
All mentions1
First seenOct 18, 2026
Last mentioned12d ago

Timeline

1 articles · newest first

Halcyon trains text-to-image models without language pretraining

Authors say Halcyon-1.6B beats a JiT baseline on gFD with 275 times less pretraining compute.

Hugging Face daily papersOct 18, 20:00 UTC · 12d agoAI research3 · 1 read

Appears with

Halcyon-0.52Bai model · 1
Halcyon-1.6Bai model · 1

Entities are extracted by the model from each article. First seen Oct 18, 2026, last mentioned 12d ago. Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.