OmniLearn and OmniLearned
A framework of small, medium and large jet models meant to serve every jet task, from top tagging to anomaly detection in CMS data.
- Made by
- Vinicius Mikuni, Benjamin Nachman, with Wahid Bhimji and Chris Harris (Berkeley Lab, NERSC)
- Field
- Collider physics
- Learns from
- More than a billion jets, including ATLAS and CMS open data, JetClass and H1
- Size
- Not stated
- Open weights
- Yes
- First public
- Apr 2024
- Source
- arxiv.org/abs/2510.24066
How it relates to Flare
Trained on more than a billion real and simulated jets, with its checkpoints public. Flare is meant to go further, from jets to whole collision events.
Links
Every figure here is from the source above, checked 17 October 2026. Something changed? Email [email protected].