Datasets:
The dataset viewer is not available for this subset.
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.
SVLA Koch Pickplace V3 TsFile
This is an Apache TsFile conversion of ITHwangg/svla_koch_pickplace_v3, a LeRobot v2.1 Koch robot dataset created by team Lebotica during the LeRobot Worldwide Hackathon.
Source dataset and attribution
- Source: ITHwangg/svla_koch_pickplace_v3
- License: MIT; robot: koch; split: train; sampling rate: 30 fps
- Creator: team Lebotica. The Hugging Face repository is owned by ITHwangg; repository history lists ITHwang and ITHwangg as contributors.
- Related source: hackathon team 76 dataset
- Demo: SmolVLA hackathon demo
- No paper or formal citation is supplied by the source card.
- Task: pick a colored ball and place it in the matching colored plate. The source contains 122 episode prompts: scattered balls in episodes 0-53 and fixed-position balls in episodes 54-121.
Pinned revision b7da23338f07f53383117fa763486d3863f8749f records removal of episodes 122-125. The source README still carries a warning for those removed episodes; the current data tree contains episodes 0-121 only.
Converted files and scale
- TsFile: data/ithwangg_svla_koch_pickplace_v3.tsfile (1,092,032 bytes)
- Table: ithwangg_svla_koch_pickplace_v3
- Rows: 65,799
- Episodes/devices: 122; semantic prompts: 122
- Source shards: 122
- Source Parquet bytes: 3,367,385
- Time precision: milliseconds; range 0 to 28033 ms within episode timelines
The source meta/info.json reports 67,913 frames, but the 122 pinned Parquet files contain 65,799 rows. Conversion and validation use the actual Parquet row count.
TsFile schema
Time = round(timestamp * 1000) in milliseconds and restarts at zero for each episode.
| Role | Columns |
|---|---|
| Time | Time (TIMESTAMP) |
| TAG/device | episode_index, task_index |
| Scalar FIELD | frame_index, sample_index (renamed from index) |
| FLOAT FIELD | action[6] to action_0 through action_5 |
| FLOAT FIELD | observation.state[6] to observation_state_0 through observation_state_5 |
The six vector elements are main_shoulder_pan, main_shoulder_lift, main_elbow_flex, main_wrist_flex, main_wrist_roll, and main_gripper. The source timestamp column is dropped because it equals Time / 1000. No rows or action/state dimensions are dropped.
Encoding, videos, and alignment
- FLOAT/DOUBLE: GORILLA + LZ4
- INT32/INT64: TS_2DIFF + LZ4
- Time: TS_2DIFF + LZ4
- BOOLEAN: RLE + LZ4 (no BOOLEAN source fields occur)
- TAG: TsFile table/device mechanism
Original videos are not included. They remain under videos/chunk-000 in observation.images.side and observation.images.top, using videos/chunk-000/{video_key}/episode_{episode_index:06d}.mp4. Source meta/info.json reports 252 MP4 files. Numeric rows remain aligned through episode_index, frame_index, and meta/episodes.jsonl.
The source task_index values are only 0 and 54 despite 122 episode prompts. They are preserved as TAG values; use meta/episodes.jsonl for the natural language prompt associated with each episode.
Validation
Local validation confirmed one readable table, 65,799 rows, 122 devices, zero-starting strictly increasing Time within every episode, vector widths of six, and physical codec settings. The JSON and Markdown reports remain in the local work directory and are excluded from upload.
Usage
from tsfile import TsFileReader
reader = TsFileReader("data/ithwangg_svla_koch_pickplace_v3.tsfile")
with reader.query_table(
"ithwangg_svla_koch_pickplace_v3",
["episode_index", "task_index", "frame_index", "sample_index",
"action_0", "observation_state_0"],
batch_size=65536) as result:
print(result.read_arrow_batch().to_pandas().head())
reader.close()
- Downloads last month
- 20