Jeveration generates 512px images in 78ms via single forward pass, 122x faster
Crazy-Repeat-2006 · reddit · 2026-11-19
A Reddit post shares the paper "Jeveration: System One Models Now Generate," extending the "System One" fast parallel-decision paradigm to image generation.
Method highlights:
- Built on a frozen vision-language model, repurposing its visual-semantic knowledge for holistic image generation
- All positions are conditioned on a shared random state to coordinate parallel predictions, so each draw yields a different but coherent image
- A distribution over discrete image codes is learned first to guide single-pass generation
- One checkpoint unifies text-to-image, image-to-image, and dense prediction, each in one forward pass
Results: 512x512 images in 78ms, up to 122x faster than existing unified models, with competitive quality and dense prediction performance. Weights on Hugging Face (linlinlinlinsdada/Jeveration), code available.
Related event: Jeveration: Frozen VLM Generates 512px Images in One Forward Pass, 78ms(5 posts)→
More from Multimodal
- Building a Claude-controlled video editor over MCP: lessons from ~200 tools, approvals and context bloat — MrxAnomynous · 2026-11-20
- OpenMontage: a self-hosted pipeline for consistent AI short dramas — Direct_Grab7063 · 2026-11-20
- CapCut launches ChatGPT video plugin: generate and edit timeline-based videos in chat — alifcoder · 2026-11-20
- I gave Claude a video editor — it made a 21-shot short I can fix shot-by-shot — MrxAnomynous · 2026-11-20
- Demi Guo previews real-time AI video model for directing films like a set — ying11231 · 2026-11-20
- Full automated GTA 6 content pipeline: Apify + ElevenLabs + Claude + Codex — eptwts · 2026-11-20