← Back to project

HoloMotion vs GEAR-SONIC · Evaluation for the X2 Project

framework anatomy · the reward/termination philosophy gap · X2 integration cost · recommendation · 2026-07-23
evaluationframework
SUPERSEDED ANALYSIS: This comparison originally treated a misimplemented custom XY/3D diagnostic (38%) as official SONIC success. A historical released-default common-v4 replay scored 104/126 = 82.54%; the audited 512-motion primary 50k result is 480/512 = 93.75%. Claims below about a proven SONIC 38% ceiling or official XY-drift termination are withdrawn.

1 · What HoloMotion is

HoloMotion (Horizon Robotics, v1.4, arXiv:2605.15336) is a motion-tracking foundation stack: a sparsely-activated MoE Transformer policy (60M→0.4B params, KV-cache inference at ~300 FPS on-robot), trained with distributed PPO on 80→2000+ hours of motion, with rough-terrain + medium domain-randomization defaults. Pipeline: raw datasets → HoloSMPL (canonical human H5) → HoloRetarget (robot refs) → HDF5 → PPO. Roadmap: v1 any-pose (done, G1), v2 any-command, v3 any-terrain, v4 any-embodiment (not built).

2 · The philosophical gap that matters for us

The original 38% framework-ceiling claim was invalid; HoloMotion should be compared independently against the audited 93.75% SONIC primary baseline.
GEAR-SONICHoloMotion
Global root positionThe historical report misread anchor_pos as XY drift; official eval checks anchor Z heightReward weight 0.0; no drift termination at all
Root velocityAbsolute-error kernel, std 1.0 (a 0.15 m/s deficit is nearly free)Relative-ratio kernel (error / (|v_ref|+0.1)) — a 20% deficit costs the same at any speed
Terminationsroot height/orientation + ankle/hand height; primary removes the hidden full-XYZ foot termtimeout + fell-over (gravity projection 0.8) only
Headline metricSuccess = never violate thresholds over full clipMPJPE-global / MPJPE-local

Separate horizontal speed probes remain useful for reference-data diagnosis, but the deleted evaluator does not show that SONIC officially terminated those clips or that the policy occupied a framework-induced local optimum. HoloMotion differs in reward design and remains worth studying, but those differences must be evaluated against the audited SONIC primary baseline rather than the withdrawn 38% narrative.

3 · Retargeting layer: same family, same G1-only limitation

HoloRetarget is another Newton/Warp LM-IK retargeter, hard-coded to G1 (class HoloNewtonG1Retargeter; assets ship only unitree_g1). Porting it to X2 would face the same issues we documented with NVIDIA's soma-retargeter (straight-elbow singularity lives in the same newton.ik solver). However, it is bypassable: the robot-side H5 needs only ref_root_pos / ref_root_rot (xyzw) / ref_dof_pos — exactly what our existing GMR corpus (121k clips) and hybrid-graft pkls already contain. A pkl→H5v2 converter (~half a day) feeds our entire corpus in without touching HoloRetarget.

4 · X2 integration cost estimate

ItemWorkEstimate
Robot config YAML~530-line spec: DOF, kp/kd, action scale, limits, asset paths — all values already known from the SONIC port1–2 days
AssetsURDF + MJCF exist; USD auto-converted by IsaacLab0
Datapkl→H5v2 converter, corpus reuse0.5 day
TrainingPPO via accelerate, fits our 32-GPU setup; MoE-TF costlier per step than SONIC's MLPruns
DeploymentTheir real-robot stack is Unitree-only; X2 deployment remains ours (same as SONIC)later

Risks: young codebase (2026.04–07); embodiment transfer is a roadmap item, so X2 is uncharted there too; rough-terrain + DR defaults are a harder regime (may start flat); success-style guarantees weaker by design.

5 · Recommendation

HoloMotion may still merit an independent pilot, but the withdrawn 38% claim provides no evidence that SONIC has a framework ceiling.
  1. Write X2 robot YAML + pkl→H5 converter (≈2 days).
  2. Pilot: train motion tracking on a corpus subset (4–8 GPUs), evaluate with the frozen SONIC 512 ruler and HoloMotion metrics under explicitly matched termination semantics.
  3. Decision gate: if the pilot's displacement achievement and drift behavior beat the audited SONIC-X2 512 primary scorecard while improving matched local and global tracking errors, scale to the full corpus.
HoloMotion evaluation · auto-generated 2026-07-23