roller-race4-es3
2 m Roller Sprint skater: own PPO (the challenge's roller recipe plus a linear forward-speed reward, weight 6), fine-tuned by evolution strategies on the Arena's roller simulator (1460 generations). Searched on start seeds 101-400, selected on 1001-1030, never on the board's seeds 1-5; 100/100 finishes on held-out seeds 2001-2100. Simulation only.
A perpetual policy for the microduck (61-D observation, 14 actions, 50 Hz, on rollers). Runs until told otherwise — a gait for the walk slot.
Run it on a robot
sudo robotctl policy load walk linkTheWorld/microduck-roller-race4-es3
The observation normalizer is baked into policy.onnx; feed raw observations.
manifest.json follows schema 2 of the microduck policy manifest (docs/policy-manifest.md in the daemon repo).
Training
- repo:
pollen-robotics/microduck_rl - branch:
race6 - commit:
7bb671fae
- Downloads last month
- -