OpenBMB’s 0.9B MiniCPM-RobotTrack brings offline vision-language tracking to Unitree Go2
CyberRobooo · x · 2026-07-24
MiniCPM-RobotTrack targets offline tracking on Unitree Go2
OpenBMB’s MiniCPM-RobotTrack is a compact 0.9B end-to-end vision-language tracking model natively adapted to Unitree Go2 Edu.
- Runs fully locally and offline, reducing latency, privacy risk, and reliability issues
- Can keep tracking even after the network disconnects
- Reported to work well in dynamic and ambiguous scenarios such as patrol, escort, and disaster relief
- The benchmark chart shows it outperforming larger baselines on multiple tracking metrics
More from Embodied
- PhyAI lifts H20 inference throughput from 10 Hz to 36 Hz for Physical AI workloads — CyberRobooo · 2026-07-24
- OpenBMB opens up MiniCPM-Robot as its first embodied AI model family at WAIC — CyberRobooo · 2026-07-24
- Robot skins made by industrial knitting could enable scalable tactile sensing — mynkgoel · 2026-07-24
- Unitree shows Super Athlete AS2-W with 16 kg payload and 30+ km range — Scobleizer · 2026-07-24
- Unitree’s As2-W quadruped is shown climbing a steep rock face — blaawker · 2026-07-24
- FLUX 3 unifies image, video, audio and action prediction in one model — robrombach · 2026-07-24