The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: ValueError
Message: Expected object or value
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 291, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 364, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Column(/run_metadata/controller/workspace_x/[]) changed from number to string in row 0
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 149, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 129, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 544, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 3052, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2559, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2584, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 558, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 440, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 306, in _generate_tables
batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
examples = [ujson_loads(line) for line in original_batch.splitlines()]
~~~~~~~~~~~^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
return pd.io.json.ujson_loads(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
ValueError: Expected object or valueNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
CLab KUKA iiwa7 demonstrations: sim-source, 500 per task
Sim-source domain: 500 successful scripted-expert episodes per task, rendered by the
MuJoCo rasteriser (domain sim). Its paired target-domain set is
emopointer/clab-kuka-gaussian-target-50 (50 Gaussian-splatting episodes per task).
5000 episodes, 2656731 control steps (36.9 h at 20 Hz),
38.8 GiB of HDF5. Robot: KUKA LBR iiwa7 R800 with a Robotiq 2F-85 gripper, simulated in MuJoCo (mujoco_warp,
1 ms physics) under a Cartesian impedance controller. Every episode is a success; the task suite and its
success checks are in CLab (branch KUKA, docs/task_suite_design.md).
| Task | Tier | Instruction | Episodes | Mean length [s] | Origin |
|---|---|---|---|---|---|
| square | medium | Put the green square nut onto the red peg | 500 | 14.1 | seeds 0..499 |
| mug_hang | medium | Hang the red mug on the wooden mug tree | 500 | 19.4 | seeds 0..522 |
| color_sort | medium | Put each cube into the tray of its color | 500 | 38.1 | seeds 0..502 |
| cup_stack | medium | Nest the red cup in the green cup, then the brown cup in the red cup | 500 | 28.5 | seeds 0..499 |
| push_t | medium | Push the red T block onto the T outline | 500 | 15.8 | seeds 0..514 |
| drawer | hard | Put the green pod into the open drawer and push the drawer closed | 500 | 20.6 | seeds 0..499 |
| shape_sorter | hard | Insert all four blocks into their matching holes in the shape sorter | 500 | 61.7 | seeds 0..504 |
| gear_mesh | hard | Put the green gear onto the pin so it meshes with the purple gear | 500 | 14.9 | seeds 0..499 |
| sweep | hard | Pick up the broom and sweep the three green cubes into the square | 500 | 32.2 | seeds 0..506 |
| hook_pull | hard | Use the hook to pull the purple cube out from under the arch | 500 | 20.4 | seeds 0..508 |
Layout
README.md, audit.json
episodes.jsonl one row per episode: task, episode, seed identity, frames, bytes, instruction
<task>/selection.json selection rule, chosen episodes, failed seeds, original collection paths
<task>/<episode>/episode.h5 the episode
<task>/<episode>/episode.json result and start records (seed, world, phase, frames)
episode.h5, T rows at 20 Hz; observation row k precedes action k, no terminal image:
| Key | Shape | Meaning |
|---|---|---|
rgb/env_camera, rgb/wrist_camera |
T,180,320,3 uint8 | fixed scene camera (D455 pose) and wrist camera (D405), rectified pinhole, lossless |
trajectory/action |
T,7 | absolute TCP target in the robot base frame: x,y,z [m], rotation vector [rad], gripper (-1 closed, +1 open) |
trajectory/joint_pos, joint_vel |
T,7 | measured iiwa joints [rad, rad/s] |
trajectory/ee_pose |
T,7 | TCP from joint FK: x,y,z, qw,qx,qy,qz, base frame |
trajectory/gripper_pos |
T,1 | closure 0 open .. 1 closed |
states/* |
privileged simulator state (qpos, qvel, mocap, per-step camera extrinsics, time); not policy inputs | |
initial_state/* |
full reset state |
The file attributes metadata_json, episode_start_json and episode_result_json hold the complete collection
configuration (controller limits, camera intrinsics and extrinsics, action and observation semantics).
Seeds and reproduction
Every episode has one integer episode seed. It alone fixes the scene layout and the expert's plan noise,
whatever batch or world it was collected in. Source seeds run from 0 upwards. Each task's 500 episodes are the
first 500 successful seeds in ascending order: <task>/selection.json lists every failed seed with its failure
reason (failures) and the successes beyond the 500th (surplus).
Reproduce one episode with the CLab collector (KUKA branch):
python scripts/collect_kuka.py --task square --seeds 123 1 --num-envs 1 --domain sim --out outputs/replay
The GPU physics is not bitwise repeatable (mujoco_warp allocates contacts with atomics), so a re-run reproduces the initial scene and the plan exactly and the trajectory within float noise; the recorded states and actions are the reference.
Notes
- Scripted expert with noise 1.0 (plan variation); only successful episodes are kept.
- Images are 320x180, the size the real-robot pipeline produces after rectifying 1280x720 RealSense frames.
audit.jsonlists what was checked before upload; RGB pixels were not decoded.
Code: https://github.com/emoPointer/CLab (KUKA branch). Published by the dataset owner.
- Downloads last month
- 134