Datasets:

Ahaskar04
/

aloha-handover-no-mocap

action_a list	action_b list	box_pos list	image_overhead array 3D	image_wrist_a array 3D	image_wrist_b array 3D	phase unknown	state_a list	state_b list
[-0.0011928855690445553,-0.9549999999999998,1.165,-0.0029776615036998717,-0.295,0.002900079851628285(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163606400427492, -0.05210109330845858, 0.02996076 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-5.529156477466737e-6,-0.9599988947441068,1.1600014828166265,-6.293412373346022e-6,-0.2999805695227(...TRUNCATED)	[-3.157165194701305e-6,-0.9599999988649489,1.159999999748588,1.7468993617932007e-7,-0.29999999552803(...TRUNCATED)
[-0.015836979139098786,-0.905000000000001,1.2150000000000007,-0.052977661503699855,-0.24499999999999(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163606400427492, -0.05210109330845858, 0.027410159999999996 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.0007563570103906339,-0.9586605502914337,1.1612351335899223,-0.002154996428850395,-0.298300160053(...TRUNCATED)	[-0.00004945679281426117,-0.9599999872708811,1.1599999963190275,4.331823797114405e-6,-0.299999965761(...TRUNCATED)
[-0.02555465541266064,-0.8550000000000015,1.2650000000000008,-0.10297766150369989,-0.195000000000000(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163606400427492, -0.05210109330845858, 0.020935559999999995 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[42,70,98],[42,70,98],[42,70,98],[42,70,98],[42,70,98],[42,70,98],[42,70,98],[42,70,98],[42,70,98](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.003154849973866454,-0.9513616770639598,1.1676085797158362,-0.011139163650997801,-0.2932460545865(...TRUNCATED)	[-0.0000473979919593872,-0.9599999905449637,1.1599999963527863,6.789094887004761e-6,-0.2999999822014(...TRUNCATED)
[-0.030888918666797445,-0.805000000000002,1.315000000000001,-0.15297766150369974,-0.1450000000000004(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163606496095011, -0.05210109923188524, 0.01660117543159748 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[41,68,96],[41,68,96],[41,68,96],[41,68,96],[41,68,96],[41,68,96],[41,68,96],[41,68,96],[41,68,96](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.006674501732369111,-0.9356021037380253,1.1814150447730556,-0.028370579316039916,-0.2838944765196(...TRUNCATED)	[-0.000011828424165002847,-0.9599999981281572,1.1599999985161669,4.127924371374214e-6,-0.30000000241(...TRUNCATED)
[-0.031620491047339466,-0.7550000000000032,1.365000000000001,-0.2029776615036997,-0.0949999999999995(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163607149361872, -0.05210115956281794, 0.017737469416642535 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[39,66,92],[39,66,92],[39,66,92],[39,66,92],[39,66,92],[39,66,92],[39,66,92],[39,66,92],[39,66,92](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.010487572984208278,-0.9110396184180038,1.203206238017497,-0.053312783071452347,-0.26987079567373(...TRUNCATED)	[9.385162689939412e-6,-0.9600000008186993,1.1599999997215873,3.449999334485963e-7,-0.300000007372859(...TRUNCATED)
[-0.027001391664994538,-0.7050000000000044,1.4150000000000014,-0.25297766150369966,-0.04499999999999(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163607745676779, -0.052101240554381326, 0.01851304751485539 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[37,62,87],[37,62,87],[37,62,87],[37,62,87],[37,62,87],[37,62,87],[37,62,87],[37,62,87],[37,62,87](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.013768855176115275,-0.8785226899127225,1.2324662495792318,-0.08472077742653504,-0.25121445895764(...TRUNCATED)	[0.000015155829055512,-0.9600000010144614,1.160000000056829,-2.0732350843864466e-6,-0.30000000652256(...TRUNCATED)
[-0.01453519966389278,-0.6550000000000042,1.4650000000000019,-0.29221380014384946,0.0050000000000006(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163607993357238, -0.0521012870163039, 0.01883612517352125 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[34,58,81],[34,58,81],[34,58,81],[34,58,81],[34,58,81],[34,58,81],[34,58,81],[34,58,81],[34,58,81](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.015710308858240055,-0.839471303188398,1.268125267283376,-0.1210788376662412,-0.22823174263148224(...TRUNCATED)	[0.000014329265607232041,-0.9600000007327599,1.1600000000792583,-2.9537557481764115e-6,-0.3000000050(...TRUNCATED)
[0.0024189902529349614,-0.6050000000000039,1.4902210007918995,-0.24221380014384966,0.055000000000000(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.05163608094525548, -0.052101306648840634, 0.01895838101190637 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[32,53,74],[32,53,74],[32,53,74],[32,53,74],[32,53,74],[32,53,74],[32,53,74],[32,53,74],[32,53,74](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.01527744254132212,-0.7954551275286285,1.3085354437466428,-0.15655292951602512,-0.201108444371893(...TRUNCATED)	[0.000012017560722820843,-0.9600000005469971,1.160000000048218,-2.8960299923836244e-6,-0.30000000403(...TRUNCATED)
[0.014504987731121709,-0.5550000000000034,1.4716830159072727,-0.1922138001438495,0.10500000000000109(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.051636081340495854, -0.052101314322927715, 0.01900615253997844 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[28,48,67],[28,48,67],[28,48,67],[28,48,67],[28,48,67],[28,48,67],[28,47,67],[28,47,66],[28,47,66](...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.012058714050597746,-0.7478226789618834,1.348906614779317,-0.18109121619329827,-0.169280825138064(...TRUNCATED)	[9.905585628330024e-6,-0.9600000004832228,1.1600000000281632,-2.4434655495307088e-6,-0.3000000034326(...TRUNCATED)
[0.017897864611126427,-0.5050000000000032,1.426751760828928,-0.14221380014384946,0.15500000000000103(...TRUNCATED)	[ 0, -0.96, 1.16, 0, -0.3, 0, 0.002 ]	[ 0.0516360814991746, -0.05210131738219965, 0.01902540284812703 ]	[[[23,38,53],[23,38,53],[22,38,53],[22,38,53],[22,38,53],[22,37,53],[22,37,53],[22,37,52],[22,37,52](...TRUNCATED)	[[[23,47,70],[23,47,71],[23,48,71],[23,48,71],[24,48,71],[28,51,75],[38,61,83],[54,77,100],[51,74,99(...TRUNCATED)	[[[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,70,99],[42,71,99](...TRUNCATED)	[ 97, 112, 112, 114, 111, 97, 99, 104 ]	[-0.0073237880994153765,-0.6977710832601776,1.3822812664390713,-0.19127011478627168,-0.1331441341053(...TRUNCATED)	[8.24124429188767e-6,-0.9600000004753046,1.1600000000205588,-1.9060000485424088e-6,-0.30000000302873(...TRUNCATED)

End of preview. Expand in Data Studio

ALOHA 2-Arm Handover (no mocap markers)

163 successful scripted demonstrations of a bimanual box handover in MuJoCo, recorded on the ALOHA / ViperX 300 dual-arm platform. Arm A picks a box off the table and hands it to arm B, which retracts with it.

Why "no mocap"

This is a re-collection of Ahaskar04/aloha-handover-data with a visual confound removed.

The scripted expert drives the arms through MuJoCo mocap bodies (left/target, right/target) — 2 cm spheres that are moved to each phase's IK goal, with the arm following after. In the original dataset those spheres were rendered into every camera frame, occupying ~6% of the wrist view (≈3,800 of 65,536 pixels) and moving to exactly where the arm was about to go.

That gives a behaviour-cloning policy an easy shortcut: track the green sphere instead of reasoning about the box and the arm's own pose. The shortcut breaks at inference, because nothing sets mocap_pos during a rollout — the spheres sit frozen at their reset position, so the policy's strongest cue simply stops moving.

Here the marker geoms are set to rgba="0 1 0 0" (alpha 0). Verified: moving a mocap body changes 0 pixels in all three cameras, against ~36,000 before.

Each episode_XXXX.h5 holds one episode. T varies per episode (mean 155, range 148–163); 25,290 timesteps in total.

Key	Shape	Dtype	Description
`image_wrist_a`	(T, 256, 256, 3)	uint8	arm A wrist camera
`image_wrist_b`	(T, 256, 256, 3)	uint8	arm B wrist camera
`image_overhead`	(T, 256, 256, 3)	uint8	fixed third-person view
`state_a`	(T, 7)	float64	arm A measured joint positions (6 joints + gripper)
`state_b`	(T, 7)	float64	arm B measured joint positions
`action_a`	(T, 7)	float64	arm A commanded targets (6 joint angles + gripper)
`action_b`	(T, 7)	float64	arm B commanded targets
`box_pos`	(T, 3)	float64	box xyz (ground truth; not available at inference)
`phase`	(T,)	S20	phase label (ground truth; useful for diagnostics)

Episode attribute: success (bool) — all episodes here are successes.

Action space

Indices 0–5 are joint angles in radians: waist, shoulder, elbow, forearm_roll, wrist_angle, wrist_rotate. Index 6 is the gripper, which takes only two values: 0.002 (closed) or 0.037 (open) — the actuator's ctrlrange.

Treating the gripper as continuous is a trap. Its range spans 0.035 while the joints span 0.11–0.91 in standard deviation, so under a uniformly averaged L1 loss it contributes ~2% of the gradient and is effectively ignored. It is better handled as binary classification.

Phases

approach → open_gripper → descend → grasp → lift
         → move_arm_B → approach_b → grip_b → release_a → retract_b

descend and grasp are the shortest phases (~10 and ~2 steps) and carry the largest joint motion, making them the hardest transitions to imitate.

Recording details

Control rate 50 Hz — the simulator runs at 500 Hz (timestep = 0.002) and one frame is recorded every 10 physics steps. A policy replaying these actions must hold each one for 10 steps to match.
Box spawn — uniform in x ∈ [-0.12, 0.12], y ∈ [-0.10, 0.10], at z = 0.03. Evaluating outside this range is out of distribution.
Initial pose — the neutral_pose keyframe, i.e. mj_resetDataKeyframe, not mj_resetData (which leaves every joint at zero, ~1.16 rad away at the elbow).
Expert is an IK-based scripted policy (mink + DAQP).
163 successes out of 200 attempts; failures were discarded.

Loading

import h5py
from huggingface_hub import snapshot_download

path = snapshot_download("Ahaskar04/aloha-handover-no-mocap", repo_type="dataset")

with h5py.File(f"{path}/episode_0000.h5", "r") as f:
    images = f["image_wrist_a"][:]   # (T, 256, 256, 3) uint8
    actions = f["action_a"][:]       # (T, 7) float64

Known limitations

Expert trajectories only. Every state lies on a successful IK path, so there are no examples of recovering from a mistake. Policies trained by plain behaviour cloning drift off-distribution during closed-loop rollout and have no demonstration to fall back on.
Single task, single object, fixed camera placement.
163 episodes is small for this setting.

Downloads last month: 190

Number of rows:

3,728

Estimated number of rows:

24,267

Total file size:

14.9 GB