Armory serves VLA policies to 14 robots from one cloud GPU at 30Hz, accepted at CoRL 2026
danfei_xu · x · 2026-09-08
Armory, a system for serving large robot policies (VLA/WAM) from cloud/edge to a fleet of robots, has been accepted at CoRL 2026 in Austin. It schedules action chunks in real time, boosting task throughput by up to 18% over baselines on heterogeneous real-world fleets. A demo shows a 'LEGO sorting factory' with 14 robots scheduled onto a single cloud GPU at 30Hz using train-time RTC.
More from Embodied
- GPT-6 Astra vs MediaPipe for hand pose: SOTA reasoning, 3 min per frame — chris_j_paxton · 2026-09-08
- After 10+ Kindles, this reader says the Xteink x3 is the most portable yet — 6 books in 23 days — Technovangelist · 2026-09-08
- Lenovo's AI PC share hits 25.1% globally as exec says cloud token costs drive adoption — 智东西 · 2026-09-08
- UW and Ai2 unveil Flex-π, a 6B-parameter robot policy that can repair its own gripper — rbhar90 · 2026-09-08
- Waymo Now Runs Fully Driverless in Austin While SF Rides Still Have Safety Drivers — sloppenheimer · 2026-09-08
- Wuji open-sources MINT, turning 1,000+ hours of egocentric video into robot training data — CyberRobooo · 2026-09-08