Hugging Face's logo

Datasets:
azithromycin
/
eastflag_g1_can_paperbox

Tasks:
Robotics
Modalities:
Time-series
Tabular
Size:
10K<n<100K
Tags:
tsfile
timeseries
tabular
robotics
lerobot
License:
Dataset card Files
xet
 Community
Dataset Viewer
The dataset viewer is not available for this subset.
Cannot get the split names for the config 'default' of the dataset.
Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "tsfile/tsfile_py_cpp.pyx", line 567, in tsfile.tsfile_py_cpp.tsfile_reader_new_c
              tsfile.exceptions.FileOpenError: 28: 
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                                         ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/tsfile/tsfile.py", line 271, in _split_generators
                  scan = self._scan_metadata(all_files)
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/tsfile/tsfile.py", line 318, in _scan_metadata
                  with self._open_reader(file) as reader:
                       ~~~~~~~~~~~~~~~~~^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/tsfile/tsfile.py", line 742, in _open_reader
                  return TsFileReader(file)
                File "tsfile/tsfile_reader.pyx", line 323, in tsfile.tsfile_reader.TsFileReaderPy.__init__
              SystemError: <class '_weakrefset.WeakSet'> returned a result with an exception set
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 71, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                               ~~~~~~~~~~~~~~~~~~~~~~~^
                      path=dataset,
                      ^^^^^^^^^^^^^
                      config_name=config,
                      ^^^^^^^^^^^^^^^^^^^
                      token=hf_token,
                      ^^^^^^^^^^^^^^^
                  )
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
                  info = get_dataset_config_info(
                      path,
                  ...<6 lines>...
                      **config_kwargs,
                  )
                File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

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.

G1 Can Paperbox TsFile

This dataset is an Apache TsFile conversion of eastflag/g1_can_paperbox, a LeRobot v2.1 robot dataset recorded with a Unitree G1 Inspire platform.

Source Dataset

  • Source: eastflag/g1_can_paperbox
  • Publisher/author: Dongkee Lee (eastflag)
  • Repository contributors: one (eastflag), as shown by the source file history
  • License: Apache-2.0
  • Source task label: g1_can_paperbox
  • Paper/citation: the source card does not provide a paper and its BibTeX field is [More Information Needed]
  • Split: train, episodes 0:69
  • Scale: 69 episodes, 34,391 frames, one task, 30 fps
  • Source frame files: 69 Parquet episode shards in data/chunk-000/
  • Source videos: 276 MP4 files (four streams for each episode)

The source card does not include a prose task description beyond the g1_can_paperbox task label, so no additional task semantics are inferred here.

Converted Artifact

  • TsFile: data/g1_can_paperbox.tsfile
  • Table: g1_can_paperbox
  • Rows: 34,391
  • Files: one merged TsFile for the train split
  • Time precision: milliseconds
  • Size: 5,024,384 bytes, or 59.31% of the 8,471,808-byte source Parquet total

Schema

Column/group TsFile role Type Description
Time TIME TIMESTAMP/INT64 round(timestamp * 1000) milliseconds; restarts at 0 for every episode
episode_index TAG STRING device segment Preserved source episode index, values 0-68
task_index TAG STRING device segment Preserved source task index, value 0
frame_index FIELD INT64 Frame number within the episode
sample_index FIELD INT64 Source index, renamed to avoid ambiguity
observation_state_0 ... observation_state_25 FIELD FLOAT Flattened observation.state[26]
action_0 ... action_25 FIELD FLOAT Flattened action[26]

The 26 state/action elements follow the source order: left arm joints (7), right arm joints (7), left Inspire hand joints (6), then right Inspire hand joints (6). The exact names remain available in meta/info.json under tsfile_conversion.source_vector_features.

Conversion Details

  • Time = round(timestamp * 1000); the source timestamp is dropped because it equals Time / 1000 seconds.
  • frame_index, all 52 numeric state/action elements, all rows, episodes, and task values are preserved.
  • Source index is renamed to sample_index.
  • Rows are sorted by episode_index, task_index, then Time.
  • FLOAT/DOUBLE use GORILLA + LZ4.
  • INT32/INT64 and Time use TS_2DIFF + LZ4.
  • BOOLEAN would use RLE + LZ4; this dataset has no BOOLEAN fields.
  • episode_index and task_index use the TsFile table/device TAG mechanism.
  • The physical codecs and all 34,391 rows were read back with the Apache TsFile Java reader.

Videos

Videos are not included in this TsFile upload. They remain in the original dataset at videos/chunk-000/ with this layout:

Each directory contains episode_000000.mp4 through episode_000068.mp4. Use the preserved episode_index and frame_index to align TsFile rows with the corresponding 30 Hz source video frames.

Read Example 

from tsfile import ColumnCategory, TsFileReader

path = "data/g1_can_paperbox.tsfile"
reader = TsFileReader(path)
schema = reader.get_all_table_schemas()["g1_can_paperbox"]
columns = [
    column.get_column_name()
    for column in schema.get_columns()
    if column.get_category() in (ColumnCategory.TAG, ColumnCategory.FIELD)
]

with reader.query_table("g1_can_paperbox", columns, batch_size=4096) as result:
    batch = result.read_arrow_batch()
    print(batch.slice(0, 5).to_pandas())

Attribution

Downloads last month
6
Total file size:
5.46 MB