README.md

metadata

license: apache-2.0
task_categories:
  - robotics
tags:
  - LeRobot
  - robotics
  - tsfile
  - timeseries
  - format:tsfile
modality:
  - tabular
  - timeseries
pretty_name: Piper Stacking Realigned V2 (TsFile)
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/piper_stacking_realigned_v2.tsfile
size_categories:
  - 10K<n<100K

Piper Stacking Realigned V2 (TsFile)

This is a time-series conversion of axiboai/piper_stacking_realigned_v2, a LeRobot v2.1 robotics dataset containing cube-stacking demonstrations from a bimanual Piper robot.

• Modalities: Time-series
• Robot: piperx_bimanual
• Source task labels: stack red cube and stack on blue cube and stack red cube on blue cube
• License: Apache License 2.0

Dataset scale

Split	Episodes	Frames / TsFile rows	Sampling rate	Tasks	Source Parquets	Final TsFiles
train	111	38,531	30 fps	2	111	1

All numeric frame data is merged into exactly one file, data/piper_stacking_realigned_v2.tsfile, containing the table piper_stacking_realigned_v2. The file is approximately 1.79 MB.

TsFile schema

Column	TsFile role	Type	Description
Time	TIME	INT64	Milliseconds, computed as Time = round(timestamp * 1000); time restarts within each episode.
episode_index	TAG	INT64	Original episode identifier (0 through 110).
task_index	TAG	INT64	Original task identifier (0 or 1).
frame_index	FIELD	INT64	Original frame position within the episode.
sample_index	FIELD	INT64	Original global index, renamed to avoid ambiguity.
observation_state_0 ... observation_state_13	FIELD	FLOAT (float32)	Complete 14-element source observation.state vector for the left and right joints and grippers.
action_0 ... action_13	FIELD	FLOAT (float32)	Complete 14-element source action vector for the left and right joints and grippers.

The source timestamp column is dropped after generating Time because it is the redundant seconds representation of the same time coordinate (Time / 1000). No rows or other numeric source columns are dropped. Both vectors are flattened without truncation, index is renamed, and the episode, task, and frame identifiers remain available.

Videos and metadata

The source contains three 480 x 640 RGB AV1 camera streams: observation.images.cam_front, observation.images.cam_left_wrist, and observation.images.cam_right_wrist. Their 333 MP4 files are intentionally omitted from this time-series repository and remain in the original dataset's videos/ tree.

Numeric rows retain episode_index, frame_index, and the 30 fps time coordinate, so they remain aligned with the corresponding source video frames. The source meta/ files are mirrored, except that meta/info.json is rewritten to describe the single TsFile path, TIME/TAG/FIELD schema, flattened and renamed features, omitted video features, and source-video alignment. No videos/ directory is included here.

Reading the data

Install the Apache TsFile Python package, then query the actual table. query_table returns the time column automatically in addition to the requested TAG and FIELD columns.

from tsfile import TsFileReader

reader = TsFileReader("data/piper_stacking_realigned_v2.tsfile")
columns = [
    "episode_index",
    "task_index",
    "frame_index",
    "sample_index",
    "observation_state_0",
    "action_0",
]

with reader.query_table(
    "piper_stacking_realigned_v2", columns, batch_size=65536
) as result:
    batch = result.read_arrow_batch()
    if batch is not None:
        print(batch.to_pandas().head())

Source and citation

The source dataset was created with LeRobot and is published at axiboai/piper_stacking_realigned_v2. Its dataset card does not provide a paper or BibTeX citation; consult the source repository for any future citation updates.