start_tick_0000
initial state before meaningful motion
Real public G1 checkpoint using the published recurrent history tensors and lower-body target mapping, driven by the staged velocity-tracking script rather than a single constant command. This page keeps the full charts, playback, and downloadable artifacts.
NVIDIA Isaac · Published G1 locomotion checkpoint on the public 29-DOF embodiment
wbc_agile
Real public G1 checkpoint using the published recurrent history tensors and lower-body target mapping, driven by the staged velocity-tracking script rather than a single constant command.
run.rrd lazily when the viewer enters the viewport.
wbc_agile
wbc_agile
assets/robots/unitree_g1/g1_29dof.xml
[2.000, 0.000, 0.000, 0.000, 0.600, 0.000, ...]
Published NVIDIA Isaac G1 ONNX checkpoint
models/wbc-agile/unitree_g1_velocity_e2e.onnx
runtime.velocity_schedule
[-0.194, 0.085, -0.032, 0.416, -0.487, -0.211, ...]
[-0.420, 0.199, -0.016, 0.691, -0.489, 0.015, ...]
Rerun recording · JSON summary · run log · Proof-pack manifest · configs/showcase/wbc_agile_real.toml
Each image overlays the target pose in blue against the actual replayed pose in orange. The checkpoints are selected from the replay trace so you can cross-check startup, motion onset, peak latency, furthest progress, and final state without opening a second report.
initial state before meaningful motion
first mean joint delta >= 0.08 rad (0.085 rad)
highest inference latency (0.791 ms)
largest planar displacement from start (2.289 m)
final recorded simulator state