THULab/aloha_mobile_chair

ALOHA Mobile Chair (TsFile)

Apache TsFile version of lerobot/aloha_mobile_chair. Modalities: Time-series.

Overview

This Mobile ALOHA dataset contains demonstrations for pushing the chairs in front of a desk until they are placed against it.

Dataset Scale

• 55 episodes, 110,000 frames, one task, sampled at 50 fps.
• The train split contains 110,000 TsFile rows in data/aloha_mobile_chair_train.tsfile.

Conversion details

Time is round(timestamp * 1000) in milliseconds. The source timestamp column is therefore not stored as a separate FIELD; frame_index is retained and source index is named sample_index. Camera videos are not included in this time-series repository and remain in the original dataset.

Schema (TsFile structure)

• Time (INT64, milliseconds) is the per-episode sample time.
• episode_index and task_index are TAG columns identifying the device series.
• frame_index, sample_index, and next_done are INT64 FIELD columns.
• observation_state_0 through _13, observation_effort_0 through _13, and action_0 through _13 are FLOAT FIELD columns for the 14 ALOHA motors.

For example, select one device with WHERE episode_index='0' AND task_index='0'.

Usage

Read data/aloha_mobile_chair_train.tsfile with the Apache TsFile Java or Python SDK.

Source & license

• Original dataset: https://huggingface.co/datasets/lerobot/aloha_mobile_chair
• Author / publisher: LeRobot / ALOHA
• Paper: https://arxiv.org/abs/2401.02117
• License: MIT