THULab/so100_413-jiangeng

SO100 413 TsFile

This dataset is converted from Jiangeng/so100_413 to Apache TsFile format.

The source dataset was created using LeRobot and contains an SO-100 robot task with synchronized video streams. This converted repository stores the numeric time-series features as Apache TsFile and keeps videos in the original Hugging Face dataset.

Modalities: Time-series.

Source Dataset

• Original dataset: Jiangeng/so100_413
• License: apache-2.0
• Task category: robotics
• Source task: Pick the cube
• LeRobot codebase version: v2.1
• Robot type: so100
• Source split: train, episodes 0:50
• Sampling rate: 30 fps
• Source videos: 100 video files across observation.images.top and observation.images.wrist_left

Converted Data

• TsFile path: data/so100_413.tsfile
• Table name: so100_413
• Converted rows: 17,851
• Episodes: 50
• Tasks: 1
• Source frame Parquet files converted: 50
• Output file size: 420,347 bytes
• Time precision: milliseconds

Schema

• Time: integer timestamp in milliseconds, computed as round(timestamp * 1000).
• TAG columns: episode_index, task_index.
• Integer FIELD columns: frame_index, sample_index.
• FLOAT FIELD columns: action_0 through action_5.
• FLOAT FIELD columns: observation_state_0 through observation_state_5.

The six action and observation-state dimensions follow the SO-100 joint names from the source metadata:

• left_shoulder_pan
• left_shoulder_lift
• left_elbow_flex
• left_wrist_flex
• left_wrist_roll
• left_gripper

Conversion Notes

• Vector columns are flattened while preserving source names: action becomes action_0..action_5, and observation.state becomes observation_state_0..observation_state_5.
• The source timestamp column is dropped because it is represented by Time / 1000 seconds after conversion.
• The source index column is renamed to sample_index.
• Video features observation.images.top and observation.images.wrist_left are omitted from this repository. Use the original dataset videos at Jiangeng/so100_413/videos.
• Source meta/ files are mirrored, with meta/info.json updated to describe the converted TsFile path and schema.

Read Example

from pathlib import Path

tsfile_path = Path("data/so100_413.tsfile")
print(tsfile_path)

# Read this file with the Apache TsFile Java/Python SDK or TsFile tools.