mcMotionChallengeGitHub ↗

MODEL / CHECKPOINT

NVIDIA

Model

Checkpoint artifacts and an independent MotionChallenge inference adapter.

Organization
NVIDIA
Type
Dense
Dense MLP motion-tracking policy.
Version
G1 BONES deploy
author checkpoint directory
Robot support
Unitree G1 · 29 DoF
Author source revision
f69fdcaabff31846b4d2247207417b57dc788828

Checkpoint artifacts

Author release
NVlabs/ProtoMotions ↗
Artifact type
Author-released checkpoint
Original file
unified_pipeline.onnx ↗

The adapter pins this published artifact by SHA256. Artifact identity is separate from inference parity.

Download links deliver the adapter files. The release links and verification notes distinguish author files from conversions; SHA256 identifies the bytes used.

Policy checkpoint
Download ↗
a59baa3e04a951e5cf0b4cc68f24ebaafa9272714226618b99a5017dfc805b4c

Inference interface

The adapter declares its control clock and reference window below. Source and configuration describe the observation and action transforms.

Observation requirements

Observation rules ↗

Model-specific handling

The actual torso orientation is FK-derived. The released pipeline takes joint state, local angular velocity, past actions and future reference torso rotations/joints; no actual translation.

Exact checkpoint & evidence ↗
Control interval
20 ms
Reference history
0 ms
Reference preview
160 ms
Output interface
See runtime contract
Configuration
policy.yaml

Evaluations