GGOSinon/robolab-det10

RoboLab det10 — deterministic scripted manipulation, successes and failures (87 tasks)

Behaviour-cloning data for training a vision-language model as a manipulation policy. Each row is (two camera images + proprioceptive state + natural language instruction) → a JSON action. No reasoning text.

Collected with a privileged scripted policy (cuRobo collision-free planning on ground-truth meshes) in RoboLab / Isaac Sim. Every attempt was kept — successes and failures.

What makes this set different: determinism

The scripted policy ran with --action-noise 0. Episodes of the same task therefore differ only by their initial state, not by injected action noise. Object placements were randomised per episode (xy offset and yaw), with a fallback ladder — 2 cm / 0.1 rad, then 1 cm / 0.05, then 0.5 cm / 0.02, then none — because the most constrained tasks reject aggressive randomisation outright ("no grasp candidates").

This is deliberate. An earlier collection replayed only 2–6 distinct initial states per task 4–41 times each, with action noise as the only variation, and 93% of its "held-out" evaluation scenes also appeared in training. Here each task was attempted 10 times from 10 distinct initial states, and the split is episode-disjoint.

Splits

split	rows	what it is
train	18052	successful episodes
eval	2246	successful episodes, held out (episode-disjoint)
failures	41891	failed attempts — not for cloning

failures is the larger half, and that is not a mistake: a failed episode runs to the full step budget while a success terminates as soon as the task predicate is satisfied, so failures are roughly 2.5× longer each. 446 failed episodes produce more rows than 627 successful ones.

Do not train a policy on failures by cloning it. These are drops, missed grasps and collisions; imitating them reproduces them. They are published because they are matched to the successes in task, scene and controller, which makes them usable for offline RL, failure prediction, reward modelling and value learning. Twenty of the 108 attempted tasks produced no successful episode and appear only here.

Action representation — and why it is deltas

The target is the next 10 end-effector displacements in millimetres, cumulative from the current pose, plus the end orientation and jaw state:

{"d_mm": [[5.0,0.6,-0.8], ..., [60.7,11.9,-15.6]],
  "quat": [0.0299, 0.9731, 0.2252, -0.0384],
  "jaw": "open"}

An earlier version used absolute TCP waypoints, and a VLM trained on it learned a degenerate solution: the end-effector sits ~530 mm from the origin while a chunk moves only ~50 mm, so echoing the input state back captures most of the token loss. The policy produced well-formed JSON and a healthy loss curve while its position error exactly equalled the error of predicting no motion. Deltas make the signal the value itself.

If you build something similar: always score against a hold-still baseline. On this data that baseline is 53.2 mm of position error, and several representations that looked healthy by loss or token accuracy scored exactly there.

Schema

column	meaning
image_overshoulder, image_wrist	PNG bytes, 392 px long side
user	instruction + state + the request string
assistant	the JSON action above
instruction	the task's natural-language goal
task, episode, step	provenance

Images are embedded in the parquet shards, so the dataset is self-contained.

Known limitations

• Per-task depth is uneven. 32 tasks have the full 10 successful episodes; 56 have fewer, and 6 have exactly one. Task breadth is good, per-task generalisation evidence is thin.
• A grasp needs millimetre precision this data does not obviously teach. A 0.8B and a 4B VLM trained on the successes both reach ~35 mm held-out position error — well above the hold-still baseline of 53.2 mm, so they have learned direction and grasp timing (96%+ jaw accuracy), but far short of what closing on an object requires. Closed-loop success on a four-task subset was 0/192 episodes across every configuration tried.
• Step 0 of every episode is near-stationary (median 0.36 mm of motion, 84.8% under 1 mm) because the scripted demonstrator has not begun moving. A policy trained on it will emit a sub-millimetre first command; if your controller rejects small commands as no-ops, that is an absorbing state.

Per-task sample counts

task	train samples
AnimalsInBinTask	135
AppleAndYogurtInBowlTask	289
BBQSauceInBinTask	47
BagelsOnPlateTask	103
BananaInBowlTask	152
BananaOnPlateTask	154
BananaThenRubiksCubeTask	329
BananasInBinOneMoreTask	276
BananasInBinThreeTotalTask	184
BananasInCrateTask	182
BananasOutOfBinTask	340
BigPumpkinInBinTask	172
BlockStackingOrderAgnosticTask	307
BlockStackingSpecifiedOrderTask	439
BlocksInBinTask	672
BowlInBinTask	242
BowlStackingLeftOnRightTask	129
BowlStackingRightOnLeftTask	103
ButterAboveRaisinTask	222
CannedFoodInBinTask	141
ClampInRightBinTask	29
ClutterPlasticTask	507
ClutterPumpkinTask	374
CookingClearPlateTask	418
CubesAndBlocksInBinTask	1334
DishesInBinTask	329
ElectronicsInBinTask	186
FoodPacking1BoxesTask	176
FoodPacking1CansTask	54
FoodPacking2BoxesTask	36
FruitsGreenLimesOnPlateTask	343
FruitsMovingOrangeOrLimeTask	168
FruitsMovingTask	241
FruitsOnPlate3Task	316
FruitsOnionTask	68
FruitsOnionToPlateTask	112
FruitsOrangesOnPlateTask	136
GrabABagelTask	118
GrabAFruitTask	129
GreenSpoonsInPotTask	288
MarkerInMugTask	242
MouseOnKeyboardTask	266
MoveBananaToBagelPlateTask	304
MustardAboveRaisinTask	122
MustardInLeftBinTask	131
MustardInRightBinTask	116
OneBottleInSquarePailTask	90
PhoneOrRemoteInBinTask	216
PickGlassesTask	110
PickOrangeObjectTask	75
PickUpBluePitcherTask	18
PickUpGreenObjectTask	10
PinkSpoonInPotTask	200
PlasticBottlesInSquarePailTask	144
PutBowlOnShelfTopTask	111
PutTwoMugsOnShelfTask	212
RecycleCartonTask	64
RedDishesInBinTask	155
RedItemsInBinTask	40
RubiksCubeAndBananaTask	259
RubiksCubeBehindBowlTask	135
RubiksCubeInFrontOfBowlTask	118
RubiksCubeLeftOfBowlTask	128
RubiksCubeOrBananaTask	118
RubiksCubeRightOfBowlTask	150
RubiksCubeTask	140
RubiksCubeThenBananaTask	278
RubiksCubesInBinTask	487
SauceBottlesCrateTask	199
SmallPumpkinInBinTask	197
SmallerObjectButterInBinTask	108
SmartphoneInBinTask	114
SpoonsInPotTask	411
Stack3RubiksCubeTask	324
StackYellowOnRedTask	141
TakeMeasuringSpoonOutTask	168
TakeSpatulaOffShelfTask	20
ThrowAwayAppleTask	149
ThrowAwaySnacksTask	362
ToolsPickingHammerTask	20
ToyInBinTask	77
UnstackRubiksCubeTask	173
UtensilsInMugTask	242
WhiteMugsInBinTask	241
WoodSpatulaToBowlTask	165
YellowAndWhiteObjectsInBinTask	384
YogurtInBowlTask	138