strata-roller-pb4

Trained and judged in Strata on microduck-challenges' roller env, own lineage: PPO on race commands with the calm-skate shaping removed and a light upright term, then four rounds of ES on the Arena roller scorer.

A perpetual policy for the microduck (61-D observation, 14 actions, 50 Hz, on rollers). Runs until told otherwise — a gait, loaded into a policy slot.

Run it on a robot

sudo robotctl policy load <slot> namenu/microduck-arena

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
Downloads last month
216
Video Preview
loading

Model tree for namenu/microduck-arena

Finetunes
8 models