wee-steps
quick steps
A perpetual policy for the microduck (61-D observation, 14 actions, 50 Hz). Runs until told otherwise — a gait, loaded into a policy slot.
Run it on a robot
sudo robotctl policy load <slot> weerawhin/microduck-wee-steps
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
- task_id:
Mjlab-HillClimb1m5-MicroDuck - repo:
https://github.com/NIRUN-Weerawit/microduck-challenges - branch:
experiments/5060ti - commit:
da3e42b5c - checkpoint:
2500 - seed:
1 - base:
mjlab-microduck 0.1.0 @ 8d0dbc94c - started:
2026-10-15T06:49:27Z
Reproduce
Same code, same uv.lock, same command, same seed. Training it again yields a comparable policy, not the same weights: GPU reinforcement learning is not bit-reproducible across machines.
git clone https://github.com/NIRUN-Weerawit/microduck-challenges
cd microduck-challenges
git checkout da3e42b5c
uv sync
uv run train Mjlab-HillClimb1m5-MicroDuck --env.scene.num-envs 2048 --agent.max-iterations 3000 --agent.seed 1 --agent.logger tensorboard --agent.run-name 30-13-30 --agent.resume True --agent.load-run '^2026\-10\-14_22\-27\-28_run\-1014\-2226$' --agent.load-checkpoint '^model_2999\.pt$'
- Downloads last month
- 15