The dataset viewer is not available for this split.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
schema: string
episodes: list<item: struct<file: string, task: string, frames: int64, duration_s: double, jitter_seed: int64, (... 38 chars omitted)
child 0, item: struct<file: string, task: string, frames: int64, duration_s: double, jitter_seed: int64, target_blo (... 26 chars omitted)
child 0, file: string
child 1, task: string
child 2, frames: int64
child 3, duration_s: double
child 4, jitter_seed: int64
child 5, target_block: string
child 6, split: string
counts: struct<total: int64, train: int64, test: int64>
child 0, total: int64
child 1, train: int64
child 2, test: int64
applied_jitter_mm: struct<>
jitter_seed: null
final_errors_m: list<item: double>
child 0, item: double
collector_success: bool
reset_block_drift_m: double
duration_s: double
placement_error_m: double
frames: int64
scout_errors_m: list<item: double>
child 0, item: double
segment_block_trace: list<item: struct<segment: string, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Le (... 37 chars omitted)
child 0, item: struct<segment: string, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Left_RightSid (... 25 chars omitted)
child 0, segment: string
child 1, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Left_RightSide_2: list<item: double>>
child 0, JengaBlock_Left: list<item: double>
child 0, item: double
child 1, JengaBlock_Left_RightSide_2: list<item: double>
...
s: struct<JengaBloc (... 77 chars omitted)
child 0, phase: string
child 1, left_s: double
child 2, best_err_mm: double
child 3, timed_out: bool
child 4, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Left_RightSide_2: list<item: double>>
child 0, JengaBlock_Left: list<item: double>
child 0, item: double
child 1, JengaBlock_Left_RightSide_2: list<item: double>
child 0, item: double
waypoints: list<item: struct<name: string, q: list<item: double>>>
child 0, item: struct<name: string, q: list<item: double>>
child 0, name: string
child 1, q: list<item: double>
child 0, item: double
output: string
smoothness: struct<max_command_joint_velocity_rad_s: double, mean_command_joint_velocity_rad_s: double>
child 0, max_command_joint_velocity_rad_s: double
child 1, mean_command_joint_velocity_rad_s: double
transfers: list<item: struct<block: string, initial: list<item: double>, destination: list<item: double>>>
child 0, item: struct<block: string, initial: list<item: double>, destination: list<item: double>>
child 0, block: string
child 1, initial: list<item: double>
child 0, item: double
child 2, destination: list<item: double>
child 0, item: double
segments: list<item: struct<name: string, duration_s: double>>
child 0, item: struct<name: string, duration_s: double>
child 0, name: string
child 1, duration_s: double
to
{'output': Value('string'), 'jitter_seed': Value('null'), 'applied_jitter_mm': {}, 'transfers': List({'block': Value('string'), 'initial': List(Value('float64')), 'destination': List(Value('float64'))}), 'scout_success': Value('bool'), 'scout_errors_m': List(Value('float64')), 'phases': List({'phase': Value('string'), 'left_s': Value('float64'), 'best_err_mm': Value('float64'), 'timed_out': Value('bool'), 'blocks': {'JengaBlock_Left': List(Value('float64')), 'JengaBlock_Left_RightSide_2': List(Value('float64'))}}), 'collector_success': Value('bool'), 'frames': Value('int64'), 'duration_s': Value('float64'), 'final_errors_m': List(Value('float64')), 'placement_error_m': Value('float64'), 'segment_block_trace': List({'segment': Value('string'), 'blocks': {'JengaBlock_Left': List(Value('float64')), 'JengaBlock_Left_RightSide_2': List(Value('float64'))}}), 'reset_block_drift_m': Value('float64'), 'segments': List({'name': Value('string'), 'duration_s': Value('float64')}), 'smoothness': {'max_command_joint_velocity_rad_s': Value('float64'), 'mean_command_joint_velocity_rad_s': Value('float64')}, 'waypoints': List({'name': Value('string'), 'q': List(Value('float64'))})}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, 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 127, 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 483, 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 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, 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 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, 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 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
schema: string
episodes: list<item: struct<file: string, task: string, frames: int64, duration_s: double, jitter_seed: int64, (... 38 chars omitted)
child 0, item: struct<file: string, task: string, frames: int64, duration_s: double, jitter_seed: int64, target_blo (... 26 chars omitted)
child 0, file: string
child 1, task: string
child 2, frames: int64
child 3, duration_s: double
child 4, jitter_seed: int64
child 5, target_block: string
child 6, split: string
counts: struct<total: int64, train: int64, test: int64>
child 0, total: int64
child 1, train: int64
child 2, test: int64
applied_jitter_mm: struct<>
jitter_seed: null
final_errors_m: list<item: double>
child 0, item: double
collector_success: bool
reset_block_drift_m: double
duration_s: double
placement_error_m: double
frames: int64
scout_errors_m: list<item: double>
child 0, item: double
segment_block_trace: list<item: struct<segment: string, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Le (... 37 chars omitted)
child 0, item: struct<segment: string, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Left_RightSid (... 25 chars omitted)
child 0, segment: string
child 1, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Left_RightSide_2: list<item: double>>
child 0, JengaBlock_Left: list<item: double>
child 0, item: double
child 1, JengaBlock_Left_RightSide_2: list<item: double>
...
s: struct<JengaBloc (... 77 chars omitted)
child 0, phase: string
child 1, left_s: double
child 2, best_err_mm: double
child 3, timed_out: bool
child 4, blocks: struct<JengaBlock_Left: list<item: double>, JengaBlock_Left_RightSide_2: list<item: double>>
child 0, JengaBlock_Left: list<item: double>
child 0, item: double
child 1, JengaBlock_Left_RightSide_2: list<item: double>
child 0, item: double
waypoints: list<item: struct<name: string, q: list<item: double>>>
child 0, item: struct<name: string, q: list<item: double>>
child 0, name: string
child 1, q: list<item: double>
child 0, item: double
output: string
smoothness: struct<max_command_joint_velocity_rad_s: double, mean_command_joint_velocity_rad_s: double>
child 0, max_command_joint_velocity_rad_s: double
child 1, mean_command_joint_velocity_rad_s: double
transfers: list<item: struct<block: string, initial: list<item: double>, destination: list<item: double>>>
child 0, item: struct<block: string, initial: list<item: double>, destination: list<item: double>>
child 0, block: string
child 1, initial: list<item: double>
child 0, item: double
child 2, destination: list<item: double>
child 0, item: double
segments: list<item: struct<name: string, duration_s: double>>
child 0, item: struct<name: string, duration_s: double>
child 0, name: string
child 1, duration_s: double
to
{'output': Value('string'), 'jitter_seed': Value('null'), 'applied_jitter_mm': {}, 'transfers': List({'block': Value('string'), 'initial': List(Value('float64')), 'destination': List(Value('float64'))}), 'scout_success': Value('bool'), 'scout_errors_m': List(Value('float64')), 'phases': List({'phase': Value('string'), 'left_s': Value('float64'), 'best_err_mm': Value('float64'), 'timed_out': Value('bool'), 'blocks': {'JengaBlock_Left': List(Value('float64')), 'JengaBlock_Left_RightSide_2': List(Value('float64'))}}), 'collector_success': Value('bool'), 'frames': Value('int64'), 'duration_s': Value('float64'), 'final_errors_m': List(Value('float64')), 'placement_error_m': Value('float64'), 'segment_block_trace': List({'segment': Value('string'), 'blocks': {'JengaBlock_Left': List(Value('float64')), 'JengaBlock_Left_RightSide_2': List(Value('float64'))}}), 'reset_block_drift_m': Value('float64'), 'segments': List({'name': Value('string'), 'duration_s': Value('float64')}), 'smoothness': {'max_command_joint_velocity_rad_s': Value('float64'), 'mean_command_joint_velocity_rad_s': Value('float64')}, 'waypoints': List({'name': Value('string'), 'q': List(Value('float64'))})}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
Jenga Phase-1 Simulation Demonstrations
Scripted, double-verified demonstrations for five tower-transfer tasks in a calibrated Isaac Sim scene (G2A Workshop Jenga Challenge, Phase 1). Every episode passed the official replay runner + strict evaluator (each block within ±3 cm / ≤15°, no collapse).
Schema jenga_nero_hdf5_v3_training, 30 Hz, per episode:
leader/qpos (N,8) commanded joints+gripper · observations/qpos (N,8) measured ·
observations/eef & leader/eef (N,7) tool pose [x y z qx qy qz qw] in robot base frame ·
observations/gripper_status (N,4) · observations/images/{left,right,wrist}_camera
(N,) jpeg 640×480. Attr stage names the per-episode jittered scene in stages/
(blocks jittered ±2 mm per seed).
Complete: 50 episodes per task, 250 total (~95 GB). Each task dir has a
manifest.json with a seed-based 45/5 train/test split.
Tasks: easy_tower_transfer_1 (2 blocks), easy_tower_transfer_2 (4, alternating layers), easy_tower_transfer_3 (6, three alternating layers), medium_tower_transfer_1 (2, cross tower), medium_tower_transfer_2 (4, double cross tower). Tooling & docs: see the companion GitHub repo.
- Downloads last month
- 654