THULab/vibepi_062626

VibePi 062626 TsFile

This dataset is a TsFile conversion of the Hugging Face dataset VibeCuisine/vibepi-062626. The source card says it was created using LeRobot. It is a LeRobot v3.0 robotics dataset for the vibeboard_follower_tilt robot type.

Modalities: Time-series. The original dataset also contains video streams, but videos are not included in this TsFile conversion.

Source Dataset

• Original dataset: VibeCuisine/vibepi-062626
• Source format: LeRobot v3.0
• License: Apache-2.0
• Robot type: vibeboard_follower_tilt
• Task: grab
• Source scale: 2 episodes, 234 frames, 1 task, 20 fps
• Source data layout: data/chunk-{chunk_index:03d}/file-{file_index:03d}.parquet
• Source video streams: observation.images.top, observation.images.wrist, and observation.images.base, each 480 x 640 AV1 video at 20 fps
• Source video files in the repository: 6

Converted Files

• data/vibepi_062626_train.tsfile
• Table name: vibepi_062626_train
• Split: train, from the source split specification 0:2
• TsFile size after conversion: 10,822 bytes
• Converted rows: 234
• Converted schema width: 23 columns total, including 1 Time column, 2 TAG columns, and 20 FIELD columns

The source meta/ directory is mirrored in this repository. meta/info.json has been updated so data_path points to the converted TsFile artifact and tsfile_conversion documents the mapping from the original LeRobot data.

Schema

• Time: integer milliseconds, computed as round(timestamp * 1000). Time restarts within each episode, matching the source per-episode timeline.
• TAG columns: episode_index, task_index.
• FIELD columns: frame_index, sample_index, scalar aux_* measurements, flattened action[7], and flattened observation.state[7].
• Source column index is renamed to sample_index.
• Source column timestamp is dropped after creating Time, because it is redundant with Time / 1000 seconds.

Vector columns are flattened into scalar TsFile fields by preserving the full source column name, replacing . with _, and appending the element index:

action[7]            -> action_0 ... action_6
observation.state[7] -> observation_state_0 ... observation_state_6

The scalar auxiliary fields are preserved as:

aux_elevator_mm
aux_spinner_active
aux_limit_home
aux_limit_gripper

Video Policy

Videos are intentionally not uploaded here. The source video features observation.images.top, observation.images.wrist, and observation.images.base remain available in the original dataset under videos/. The TsFile artifact stores the aligned numeric state/action/time/task data.

Validation

Conversion was validated locally with the project pipeline:

.\.venv\Scripts\python.exe scripts\pipeline.py inspect configs\vibepi_062626.yaml
.\.venv\Scripts\python.exe scripts\pipeline.py convert configs\vibepi_062626.yaml
.\.venv\Scripts\python.exe scripts\pipeline.py validate configs\vibepi_062626.yaml
.\.venv\Scripts\python.exe scripts\verify_tsfile.py .\workdir\vibepi_062626\tsfile\data\vibepi_062626_train.tsfile --parquet .\workdir\vibepi_062626\staged\vibepi_062626_train.parquet

The readback check opened the TsFile successfully and confirmed that the TsFile metadata row count, query row count, and staged Parquet row count all match at 234 rows.

Read Example

from tsfile import TsFileReader

path = "data/vibepi_062626_train.tsfile"
table = "vibepi_062626_train"

with TsFileReader(path) as reader:
    columns = [
        "episode_index",
        "task_index",
        "frame_index",
        "sample_index",
        "aux_elevator_mm",
        "action_0",
        "observation_state_0",
    ]
    with reader.query_table(table, columns, batch_size=4096) as result:
        batch = result.read_arrow_batch()
        print(batch)