Datasets:

azithromycin
/

nikodembartnik_rocochallenge2026_Industrial_Assembly_annotated

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.

ROCO Challenge 2026 Industrial Assembly Annotated TsFile

Apache TsFile conversion of nikodembartnik/rocochallenge2026_Industrial_Assembly_annotated, an annotated LeRobot v3 industrial-assembly dataset for a vega_1u robot.

Source, author, and license

Original dataset: https://huggingface.co/datasets/nikodembartnik/rocochallenge2026_Industrial_Assembly_annotated
Pinned source revision: 1dd33c9d939e4487c5857a9bb6001c9bbe91af8e
Original repository owner/uploader: Nikodem Bartnik (nikodembartnik)
Authorship evidence: the Hugging Face profile displays the name Nikodem Bartnik; the source card does not list a separate author group.
License: Apache-2.0
Task: assemble parts onto the task board
Paper/citation: the source card provides no paper or completed BibTeX citation.

Scale and source layout

Split: train (0:200)
200 episodes, 121,454 frame rows, one task, 10 fps
52 Parquet frame shards under data/chunk-000/
file-000.parquet through file-050.parquet each contain one episode; file-051.parquet contains episodes 51 through 199.
Source frame pattern: data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet

Converted artifact

TsFile: data/nikodembartnik_rocochallenge2026_industrial_assembly_annotated.tsfile (31,137,917 bytes)
Table: nikodembartnik_rocochallenge2026_industrial_assembly_annotated
Rows: 121,454; TAG devices: 200
All 52 source frame shards are merged into one table-model TsFile.
TsFile/source-Parquet size ratio: 87.57%
Source meta/ is retained, with meta/info.json updated for the TsFile artifact.

Schema and mapping

TsFile column/group	Role/type	Source mapping
`Time`	TIME / INT64 milliseconds	`round(timestamp * 1000)`; restarts at zero per episode
`episode_index`, `task_index`	TAG/device dimensions	Original source columns
`frame_index`	FIELD / INT64	Preserved
`sample_index`	FIELD / INT64	Renamed from `index`
`observation_state_0` ... `observation_state_43`	FIELD / FLOAT	Flattened from `observation.state[44]`
`action_0` ... `action_13`	FIELD / FLOAT	Flattened from `action[14]`
`language_persistent`, `language_events`	FIELD / STRING	Nested annotations serialized losslessly as canonical JSON text

State dimensions, in source order: left_ee_x, left_ee_y, left_ee_z, left_ee_qw, left_ee_qx, left_ee_qy, left_ee_qz, right_ee_x, right_ee_y, right_ee_z, right_ee_qw, right_ee_qx, right_ee_qy, right_ee_qz, left_jpos_0, left_jpos_1, left_jpos_2, left_jpos_3, left_jpos_4, left_jpos_5, left_jpos_6, right_jpos_0, right_jpos_1, right_jpos_2, right_jpos_3, right_jpos_4, right_jpos_5, right_jpos_6, left_jvel_0, left_jvel_1, left_jvel_2, left_jvel_3, left_jvel_4, left_jvel_5, left_jvel_6, right_jvel_0, right_jvel_1, right_jvel_2, right_jvel_3, right_jvel_4, right_jvel_5, right_jvel_6, left_gripper, right_gripper.

Action dimensions, in source order: left_ee_x, left_ee_y, left_ee_z, left_ee_rx, left_ee_ry, left_ee_rz, left_gripper, right_ee_x, right_ee_y, right_ee_z, right_ee_rx, right_ee_ry, right_ee_rz, right_gripper.

Use json.loads(...) to recover each original nested language annotation value. The source timestamp is dropped after Time synthesis because it is exactly represented by Time / 1000 seconds. No frame row, episode, task, numeric state/action dimension, or language annotation column is dropped.

Physical storage

Time: TS_2DIFF + LZ4
FLOAT/DOUBLE fields: GORILLA + ZSTD
INT32/INT64 fields: TS_2DIFF + ZSTD
BOOLEAN fields: RLE + LZ4 (the source has no BOOLEAN field)
STRING annotation fields: PLAIN + LZ4
TAG values use the TsFile table-model device/TAG mechanism.

The physical schema and codecs were inspected with the Apache TsFile Java reader after conversion.

Original videos (not included)

The source has 159 H.264 MP4 files (53 per camera stream; 1,113,405,751 bytes) at videos/{video_key}/chunk-000/file-XXX.mp4:

Files 0 through 50 contain one episode each. Video file 51 packs episodes 51-165 and file 52 packs episodes 166-199. Use the per-episode file_index, from_timestamp, and to_timestamp values in meta/episodes/, followed by frame_index at 10 fps, for exact frame alignment. Videos are not copied into or uploaded with this TsFile dataset; use the original repository links above.

Validation

Local source/staged/TsFile metadata/Java-query rows all equal 121,454. The validation also checks 200 TAG devices, zero duplicate (episode_index, task_index, Time) keys, annotation serialization hashes, and every physical codec. Local JSON and Markdown reports are intentionally excluded from this dataset directory.

Minimal read example

from tsfile import TsFileReader

reader = TsFileReader("data/nikodembartnik_rocochallenge2026_industrial_assembly_annotated.tsfile")
with reader.query_table(
    "nikodembartnik_rocochallenge2026_industrial_assembly_annotated",
    ["episode_index", "task_index", "Time", "frame_index", "action_0", "language_events"],
    batch_size=4096,
) as result:
    batch = result.read_arrow_batch()
    print(batch.to_pandas().head())
reader.close()

Please cite the original Hugging Face dataset and Nikodem Bartnik.

Downloads last month: 24

Total file size:

31.2 MB