Compsognathus Longipes
The anatomical proxy has a 1.0 kg assumed mass, 14 actuators, a long articulated tail, and 56 state observations (the last three are the Phase C body-relative command segment, zero under command_mode = "none").
Generated specifications
These values are generated from the environment and compiled MJCF model. “Dynamic mass” is a simulator property, not an anatomical weight estimate.
| Gait | Bipedal |
|---|---|
| Observation dimension | 56 |
| Action dimension / actuators | 14 |
| Generalized coordinates (nq) | 24 |
| Generalized velocities (nv) | 23 |
| Compiled dynamic model mass | 1.0 kg |
| Plant contract revisions | Policy r2 · physics r1 · visual r1 |
| Model | environments/compsognathus/assets/compsognathus.xml |
Current curriculum configuration
The Stable-Baselines3 budgets and early-advancement gates below come from the current TOML files and can differ from older published runs. A stage also ends when its configured budget is exhausted. JAX/MJX uses the same task sequence, but its CLI and notebook gate behavior is not yet equivalent. The recipe column names the behavior a stage belongs to and whether its checkpoint is a published deliverable; the warm-start column names the stage it starts from.
| Stage | Recipe | Warm-start from | Objective | SB3 configured budget | SB3 early-advancement gate |
|---|---|---|---|---|---|
| 1 — Balance | stand (deliverable) | — | Hold an upright stance with foot support | 11M | reward ≥ 1,800; full-horizon episodes ≥ 95.0%; unsupported duty ≤ 0.02; unsupported duty 95% upper bound ≤ 0.02; ≥ 40 episodes/evaluation; 3 consecutive passes |
| recovery — Recovery | stand (deliverable) | 1 — Balance | Pilot: recover an upright stance after calibrated horizontal pushes | 3M | recovery success LCB95 ≥ 0.5; paired Δ vs each required frozen null LCB95 ≥ 0.1; re-entry ≤ 40 steps + 20-step dwell; ≥ 40 episodes/evaluation; verdict from the frozen gate_resolution.json (post-stage; fail-closed when absent or stale) |
| 2 — Locomotion | walk (deliverable) | 1 — Balance | Move forward while remaining upright and avoiding body-floor contact | 3M | reward ≥ 500; episode length ≥ 900; avg. velocity ≥ 0.08 m/s; ≥ 20 episodes/evaluation; 3 consecutive passes |
| 3 — Target Reach | hunt (deliverable) | 2 — Locomotion | Reach the randomized horizontal target and slow down while upright | 3M | reward ≥ 25; task success ≥ 70.0%; ≥ 20 episodes/evaluation; 3 consecutive passes |
Backend-specific task-success semantics
- Stable-Baselines3 — Pelvis target-reaching success: While the target task is enabled, the upright pelvis enters the configured horizontal target radius at or below the configured speed; non-foot floor contact is forbidden.
Per-deliverable success semantics
- stance (1 — Balance) · Stable-Baselines3 — Stance quality (stance_quality/v1): The stance checkpoint clears the stance_quality/v1 gate: the unsupported-duty 95% upper bound and the full-horizon episode fraction meet the configured bounds over 40-episode evaluations, with the reward rail as a collapse floor only. The duty and full-horizon statistics are not exported to summary.json until a later phase (decision D-A9, deferred by D-B15), so the catalog names them with no value.
- recovery (recovery — Recovery) · Stable-Baselines3 — Recovery under pushes (recovery_quality/v1): The recovery checkpoint clears the recovery_quality/v1 gate, judged post-stage against the run's frozen gate_resolution.json: recovery-success LCB95 and the paired delta against each required frozen null meet the frozen thresholds, with re-entry inside the configured step budget and dwell. The statistics are not exported to summary.json until a later phase (decision D-A9, deferred by D-B15).
- walk (2 — Locomotion) · Stable-Baselines3 — Gated forward velocity (reward_and_length/v1): The locomotion checkpoint clears the reward_and_length/v1 gate: mean forward velocity at or above the stage's configured minimum, with its reward and episode-length floors, over the required consecutive evaluations.
- hunt (3 — Target Reach) · Stable-Baselines3 — Pelvis target-reaching success: While the target task is enabled, the upright pelvis enters the configured horizontal target radius at or below the configured speed; non-foot floor contact is forbidden.
Published run summaries
These are historical experiment records. A result is not evidence for the current model revision unless its provenance is explicitly marked current and verified.
No run summary has been published for this species.
Training
Select Compsognathus Longipes in the SB3 training notebook. Its manifest (configs/compsognathus/stages.toml) declares four nodes, all published deliverables: stance (stand), recovery (stand, warm-started from stance), locomotion (walk, warm-started from stance) and behavior (hunt — the non-contact target-reach task, warm-started from locomotion). The notebook's BEHAVIOR = "hunt" default resolves to the behavior node here as for every species. The final task requires an upright, slow arrival within 8 cm of the target, measured at the pelvis in the ground plane.
python -m pip install -e ".[train]"
python -m environments.shared.train --species compsognathus train --stage 1
# The advancing stages in manifest order; reuse an earlier run's certified stance and walk
python -m environments.shared.train --species compsognathus curriculum \
--trunk-from logs/<earlier_run> --output-dir logs/<new_run>
Actions are residual joint-position commands around a preloaded standing pose. The control rate is 50 Hz and the default episode horizon is 20 seconds. Simulator state and target coordinates are privileged MLP inputs; the RGB camera is available separately for rendering. These policies are not onboard camera controllers.
The initial configurations and short integration runs establish training compatibility. They do not establish learned walking, behavior convergence or hardware readiness. The zero-action standing baseline should be measured before tuning. Every node is judged on its own and its verdict written to gate_verdict.json beside the handoff checkpoint; a later run reuses it only under the reuse rule (a passed verdict, the same task digest and plant, chained by the parent's checkpoint digest) — see Behavior Recipes. JAX/MJX training is not implemented for this variant.
See the model and training guide for mechanical assumptions, sensor details and validation commands.