DeCAF Makes Co-folding 5x Faster
mmbronstein · x · 2026-07-10
The author released a new preprint on protein-ligand co-folding, proposing a flow-map framework named DeCAF for faster few-step co-folding inference.
The post claims that while maintaining sample quality, the method achieves up to a 5x speedup on the SOTA Pearl model and is 20x faster than Boltz 1x. Links to the blog and arXiv are provided, and the code will be released soon.
Related event: DeCAF Framework Speeds Up Protein-Ligand Co-Folding Nearly 5x(2 posts)→
More from Research
- NeurIPS 2026 workshop will focus on on-device intelligence and local execution — YiMaTweets · 2026-07-21
- NeurIPS 2026 workshop calls papers on on-device intelligence — YiMaTweets · 2026-07-21
- AI Security Institute says every tested model tried to cheat in cyber evaluations — connoraxiotes · 2026-07-21
- AI companies are buying old books to avoid training on AI-generated slop — CackleRooster · 2026-07-21
- Sakana says multiple diffusion models plus MCTS beat test-time scaling on coding and math — SakanaAILabs · 2026-07-21
- Soofi S 30B-A3B releases a full pretraining report and claims open-model leads in English and German — abursuc · 2026-07-21