Quadruped RL training in 1 minute on an M1 MacBook via mjbatch with 1024 parallel envs
philfung · x · 2026-09-12
kevinzakka open-sourced mjbatch (317 GitHub stars): MuJoCo batched parallel simulation training a Go1 quadruped to follow joystick commands across 1024 parallel environments, solving the task in just 1 minute on a 5-year-old M1 MacBook. The repo includes the full 400-line go1joystick.py training script, showing consumer laptops can handle large-scale parallel robot RL.
More from Embodied
- Smart glasses app visualizes WiFi signal strength in AR to find the best spot — Scobleizer · 2026-09-12
- Sprite humanoid robot's full FreeCAD hardware design released under CERN OHL — Obijuan_cube · 2026-09-12
- UK thinktank urges taxing robotaxis, claiming up to £46bn a year — dc_lawrence · 2026-09-12
- Rider in wheel-free Tesla Robotaxi: 'Drove better than any human' — Houston next, Musk amplifies — elonmusk · 2026-09-12
- LightParkour: humanoid robots learn whole-body parkour via Real2Sim2Real skill distillation — chris_j_paxton · 2026-09-12
- Homemade Livox LiDAR + iPad SLAM maps a cave after a 3.5-hour hike — ducha_aiki · 2026-09-12