Datasets:

rtrtwrw
/

granular-material-pour-flour-300

Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
codebase_version: string
robot_type: null
total_episodes: int64
total_frames: int64
total_tasks: int64
chunks_size: int64
data_files_size_in_mb: int64
video_files_size_in_mb: int64
fps: int64
splits: struct<train: string>
  child 0, train: string
data_path: string
video_path: string
features: struct<timestamp: struct<dtype: string, shape: list<item: int64>, names: null>, frame_index: struct< (... 1383 chars omitted)
  child 0, timestamp: struct<dtype: string, shape: list<item: int64>, names: null>
      child 0, dtype: string
      child 1, shape: list<item: int64>
          child 0, item: int64
      child 2, names: null
  child 1, frame_index: struct<dtype: string, shape: list<item: int64>, names: null>
      child 0, dtype: string
      child 1, shape: list<item: int64>
          child 0, item: int64
      child 2, names: null
  child 2, episode_index: struct<dtype: string, shape: list<item: int64>, names: null>
      child 0, dtype: string
      child 1, shape: list<item: int64>
          child 0, item: int64
      child 2, names: null
  child 3, index: struct<dtype: string, shape: list<item: int64>, names: null>
      child 0, dtype: string
      child 1, shape: list<item: int64>
          child 0, item: int64
      child 2, names: null
  child 4, task_index: struct<dtype: string, shape: list<item: int64>, names: null>
      child 0, dtype: string
      child 1, shape: list<item: int64>
          child 0, item: int64
      child 2, names: null
  child 5, observation.stat
...

      child 3, info: struct<video.height: int64, video.width: int64, video.codec: string, video.pix_fmt: string, video.is (... 75 chars omitted)
          child 0, video.height: int64
          child 1, video.width: int64
          child 2, video.codec: string
          child 3, video.pix_fmt: string
          child 4, video.is_depth_map: bool
          child 5, video.fps: int64
          child 6, video.channels: int64
          child 7, has_audio: bool
  child 9, action: struct<dtype: string, shape: list<item: int64>, names: list<item: string>>
      child 0, dtype: string
      child 1, shape: list<item: int64>
          child 0, item: int64
      child 2, names: list<item: string>
          child 0, item: string
  child 10, is_teleop: struct<dtype: string, shape: list<item: int64>, names: null>
      child 0, dtype: string
      child 1, shape: list<item: int64>
          child 0, item: int64
      child 2, names: null
station_config: struct<stand_name: string, robot_type: string, vcodec: string, configured_at: int64>
  child 0, stand_name: string
  child 1, robot_type: string
  child 2, vcodec: string
  child 3, configured_at: int64
checkpoint_metadata: null
loop_rate_hz: int64
git_commit_hash: string
rtc_metadata: null
task_name: string
hostname: string
eval_run_id: null
teleoperator_name: string
task_description: null
creation_time_epoch_ns: int64
episode_outcome: null
yam_calibration: null
policy_type: null
eval_run_date: null
collection_method: string
git_dirty: bool
to
{'hostname': Value('string'), 'git_commit_hash': Value('string'), 'git_dirty': Value('bool'), 'creation_time_epoch_ns': Value('int64'), 'task_name': Value('string'), 'task_description': Value('null'), 'teleoperator_name': Value('string'), 'collection_method': Value('string'), 'episode_outcome': Value('null'), 'station_config': {'stand_name': Value('string'), 'robot_type': Value('string'), 'vcodec': Value('string'), 'configured_at': Value('int64')}, 'yam_calibration': Value('null'), 'checkpoint_metadata': Value('null'), 'rtc_metadata': Value('null'), 'loop_rate_hz': Value('int64'), 'policy_type': Value('null'), 'eval_run_id': Value('null'), 'eval_run_date': Value('null')}
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 2951, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2461, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ~~~~~~~~~~~~~~~~^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2486, 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 547, in _iter_arrow
                  for key, pa_table in iterator:
                                       ^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 430, 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
              codebase_version: string
              robot_type: null
              total_episodes: int64
              total_frames: int64
              total_tasks: int64
              chunks_size: int64
              data_files_size_in_mb: int64
              video_files_size_in_mb: int64
              fps: int64
              splits: struct<train: string>
                child 0, train: string
              data_path: string
              video_path: string
              features: struct<timestamp: struct<dtype: string, shape: list<item: int64>, names: null>, frame_index: struct< (... 1383 chars omitted)
                child 0, timestamp: struct<dtype: string, shape: list<item: int64>, names: null>
                    child 0, dtype: string
                    child 1, shape: list<item: int64>
                        child 0, item: int64
                    child 2, names: null
                child 1, frame_index: struct<dtype: string, shape: list<item: int64>, names: null>
                    child 0, dtype: string
                    child 1, shape: list<item: int64>
                        child 0, item: int64
                    child 2, names: null
                child 2, episode_index: struct<dtype: string, shape: list<item: int64>, names: null>
                    child 0, dtype: string
                    child 1, shape: list<item: int64>
                        child 0, item: int64
                    child 2, names: null
                child 3, index: struct<dtype: string, shape: list<item: int64>, names: null>
                    child 0, dtype: string
                    child 1, shape: list<item: int64>
                        child 0, item: int64
                    child 2, names: null
                child 4, task_index: struct<dtype: string, shape: list<item: int64>, names: null>
                    child 0, dtype: string
                    child 1, shape: list<item: int64>
                        child 0, item: int64
                    child 2, names: null
                child 5, observation.stat
              ...
              
                    child 3, info: struct<video.height: int64, video.width: int64, video.codec: string, video.pix_fmt: string, video.is (... 75 chars omitted)
                        child 0, video.height: int64
                        child 1, video.width: int64
                        child 2, video.codec: string
                        child 3, video.pix_fmt: string
                        child 4, video.is_depth_map: bool
                        child 5, video.fps: int64
                        child 6, video.channels: int64
                        child 7, has_audio: bool
                child 9, action: struct<dtype: string, shape: list<item: int64>, names: list<item: string>>
                    child 0, dtype: string
                    child 1, shape: list<item: int64>
                        child 0, item: int64
                    child 2, names: list<item: string>
                        child 0, item: string
                child 10, is_teleop: struct<dtype: string, shape: list<item: int64>, names: null>
                    child 0, dtype: string
                    child 1, shape: list<item: int64>
                        child 0, item: int64
                    child 2, names: null
              station_config: struct<stand_name: string, robot_type: string, vcodec: string, configured_at: int64>
                child 0, stand_name: string
                child 1, robot_type: string
                child 2, vcodec: string
                child 3, configured_at: int64
              checkpoint_metadata: null
              loop_rate_hz: int64
              git_commit_hash: string
              rtc_metadata: null
              task_name: string
              hostname: string
              eval_run_id: null
              teleoperator_name: string
              task_description: null
              creation_time_epoch_ns: int64
              episode_outcome: null
              yam_calibration: null
              policy_type: null
              eval_run_date: null
              collection_method: string
              git_dirty: bool
              to
              {'hostname': Value('string'), 'git_commit_hash': Value('string'), 'git_dirty': Value('bool'), 'creation_time_epoch_ns': Value('int64'), 'task_name': Value('string'), 'task_description': Value('null'), 'teleoperator_name': Value('string'), 'collection_method': Value('string'), 'episode_outcome': Value('null'), 'station_config': {'stand_name': Value('string'), 'robot_type': Value('string'), 'vcodec': Value('string'), 'configured_at': Value('int64')}, 'yam_calibration': Value('null'), 'checkpoint_metadata': Value('null'), 'rtc_metadata': Value('null'), 'loop_rate_hz': Value('int64'), 'policy_type': Value('null'), 'eval_run_id': Value('null'), 'eval_run_date': Value('null')}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

granular-material-pour -- flour, single target

122 teleoperated pours on a single YAM arm. 2.80 hours, 50 Hz, 3 cameras and a 1 kg load cell. Instruction: pour flour onto the scale.

Nominally a single-target session at about 300 g.

Schema

Feature	Shape
`observation.state`	`[9]`	`left_joint_1..6.pos`, `left_gripper.pos`, `scale_remaining.g`, `scale_target.g`
`action`	`[7]`	`left_joint_1..6.pos`, `left_gripper.pos`
`observation.images.back`	video	640x480 AV1, 50 fps
`observation.images.top`	video	640x480 AV1, 50 fps
`observation.images.wrist_left`	video	640x480 AV1, 50 fps

scale_remaining.g = scale_target.g - the weight in the pan. It counts down to zero and goes negative on an overshoot, so its sign is the instruction: positive means add, negative means take out, near zero means stop.

Stored targets span 296-309 g across 105 distinct values.

Things worth knowing before you train

Three episodes run backwards. They start near 303 g and end near zero: the material is taken off the scale, under the same instruction pour flour onto the scale that every other episode carries. One sentence covers two opposite behaviours and nothing in the data separates them.
scale_target.g is a sensor reading. One nominal target became 105 distinct stored values spanning 296-309 g.

Where this sits in the collection

This release holds eight sessions and three incompatible schemas. Six sessions record at 50 Hz with three cameras; two record at 30 Hz with four, the fourth aimed at the scale display. Three of the 50 Hz sessions are 8 wide and carry no live weight. A LeRobot dataset holds one schema, so these are separate corpora, not one -- which is why each session is published as its own repository.

Every defect we found is counted and located in the accompanying paper.

Downloads last month: 362

Total file size:

24.8 GB

Collection including rtrtwrw/granular-material-pour-flour-300

Weight pouring dataset

Collection

The collection of datasets used fir training the root pouring weights, 1g precigion • 9 items • Updated 5 days ago