⚠ SUPERSEDED — do not deploy
The current, tested deliverables are:
- π0.5:
oku006/bussing-pi05-droid via branch pi05-joint-velocity of sombit888/robot_evals
- whiteboard:
oku006/bussing-whiteboard-act via branch whiteboard-joint-velocity
This repository is kept only as history/provenance. (21 Aug 2026)
license: apache-2.0
tags: [robotics, act, vla, insait]
pipeline_tag: robotics
INSAIT whiteboard policy — gripper-weighted variant
A variant of oku006/bussing-whiteboard-act,
not a separate model. Same architecture, same data, one training change: the
gripper channel is normalized alongside the seven velocity channels instead of
being left raw.
Why it exists: left raw (std 0.27) against normalized velocities (std 1.0), the
gripper received only ~3.7% of the L1 signal. Normalizing it gives it a fair
1/8. The effect is a clean trade — gripper error improves 20% (0.0266 →
0.0212) and clipping drops from 2.9% to 0.5%, while the velocity metrics get
slightly worse — because the total loss signal is conserved.
Use the main checkpoint by default. Reach for this one if the arm reaches
the right place but the grasp closes at the wrong moment, which is the specific
failure this addresses.
1from bussing_act import BussingActPolicy # shipped in this repo
2policy = BussingActPolicy.from_pretrained(".", device="cuda")
3out = policy.infer({
4 "observation/image": main_uint8_hwc, # overhead camera
5 "observation/wrist_image": wrist_uint8_hwc, # wrist camera
6 "observation/state": joints_and_gripper # 8 floats
7})
8out["actions"] # (16, 8) = 7 joint velocities rad/s + gripper
Loads with no training repository and no network access — the frozen encoder
is rebuilt from a config stored here, not downloaded.
Architecture
| |
|---|
| vision encoder | frozen DINOv2-base ViT (86.6M), swappable |
| decoder | 7-layer transformer decoder, learned action queries |
| encoder | 4-layer, d_model 512, 8 heads, FFN 3200 |
| CVAE | off |
| language conditioning | supported (lang_mode token / FiLM); unused here, single task |
| parameters | 142.3M total, 55.7M trainable |
| chunk | 16 steps, first 8 executed |
The frozen swappable encoder is the point of the project: the vision backbone is
pretrained and never updated, so the policy learns the task rather than the
representation, and SigLIP / DINOv2 / JEPA can be exchanged without touching
anything else. A SigLIP variant of this checkpoint also exists.
Action space
7 joint velocities (rad/s) + gripper position, matching
JointVelocitiesAction(velocities=action[:7]). No conversion step, no
end-effector pose needed, no workspace clamp — a velocity is not a position.
Set flip_gripper: False: trained on the raw recorded convention
(1.0 = open, ~0.3 = closed on an object).
Measured — held-out episodes [2, 4, 10, 20, 21], 8 executed steps
Identical protocol for every row; the shared zero-motion baseline of
0.02671 rad/s is the check that they are comparable.
| π0.5 | π0.5 own-stats | main whiteboard | this (gripper-weighted) | LeRobot ACT |
|---|
| beats zero-motion | 29.9% | 28.9% | 27.4% | 25.5% | 24.1% |
| velocity error (rad/s) | 0.0187 | 0.0190 | 0.0194 | 0.0199 | 0.0203 |
| gripper error | 0.0073 | 0.0097 | 0.0266 | 0.0212 | 0.0349 |
| motion ratio (1.0 = human) | 0.921 | 0.928 | 0.857 | 0.845 | 0.940 |
| joint error, 8 steps | 0.0286 | 0.0289 | 0.0296 | 0.0304 | 0.0307 |
| clipped by controller | 1.8% | 2.8% | 2.9% | 0.5% | 2.1% |
At 142M parameters with no robot pretraining, this sits within 2.5 points of a
3B model pretrained on thousands of hours of DROID data, and ahead of LeRobot's
stock ACT. Inference 15 ms.
Limits
33 of 38 episodes for training, 5 held out. One task. No generalization to new
objects or layouts expected. Every recording starts from ≈ the same home pose
(1 cm spread), so starting far from it is out of distribution. Best validation
came at step 13,000 of 20,000 — it overfits after that, which is why the
best-validation checkpoint is the one shipped.
