pi0_test_0205 (TsFile)

Apache TsFile version of Tank-123/pi0_test_0205.

Overview

A LeRobot teleoperation dataset recorded on an SO-101 follower arm. Each episode is one demonstration of a single manipulation task — "Grab the red box." — captured at 30 fps with 6-DOF joint position observations and actions.

• Robot: SO-101 follower (6-DOF arm).
• Scale: 72 episodes, 22,938 frames total, 1 task, recorded at 30 fps.
• Per frame: 6-dimensional action and observation.state (joint positions: shoulder_pan, shoulder_lift, elbow_flex, wrist_flex, wrist_roll, gripper).
• Cameras: two video streams (laptop, phone, 480×640) in the original dataset — see the note below; not included here.

Schema (TsFile structure)

All 72 episodes are stored in a single .tsfile using the TsFile table model.

• Time (INT64, milliseconds) — per-frame timestamp, round(timestamp * 1000), restarting at 0 for each episode.
• episode_index (TAG) — episode identifier (0–71). Query one episode with WHERE episode_index=0.
• task_index (TAG) — task identifier (always 0 here; the task is "Grab the red box.").
• frame_index (FIELD, INT64) — frame number within the episode.
• sample_index (FIELD, INT64) — the original global index column, renamed.
• action_0 .. action_5 (FIELD, FLOAT) — 6-DOF commanded joint positions.
• observation_state_0 .. observation_state_5 (FIELD, FLOAT) — 6-DOF observed joint positions.

The fixed-length vector columns action and observation.state are flattened to action_0..5 and observation_state_0..5 (single-precision FLOAT). The redundant source timestamp column is dropped (it equals Time ÷ 1000 seconds).

Usage

Read the .tsfile files with the Apache TsFile Java or Python SDK.

Source & license

• Original dataset: https://huggingface.co/datasets/Tank-123/pi0_test_0205
• Author / publisher: Tank-123
• Created with: LeRobot (codebase v2.1)
• Camera videos (observation.images.laptop, observation.images.phone) are not included in this repository; see the videos/ directory of the original dataset.
• License: apache-2.0