MS-HAB-TidyHouse / README.md
arth-shukla's picture
Update README.md
1f303a4 verified
metadata
language:
  - en
license: mit
tags:
  - robotics
  - manipulation
  - rearrangement
  - computer-vision
  - reinforcement-learning
  - imitation-learning
  - rgbd
  - rgb
  - depth
  - low-level-control
  - whole-body-control
  - home-assistant
  - simulation
  - maniskill
annotations_creators:
  - machine-generated
language_creators:
  - machine-generated
language_details: en-US
pretty_name: ManiSkill-HAB TidyHouse Dataset
size_categories:
  - 1M<n<10M
task_categories:
  - robotics
  - reinforcement-learning
task_ids:
  - grasping
  - task-planning
configs:
  - config_name: pick-002_master_chef_can
    data_files:
      - split: trajectories
        path: pick/002_master_chef_can.h5
      - split: metadata
        path: pick/002_master_chef_can.json
  - config_name: pick-003_cracker_box
    data_files:
      - split: trajectories
        path: pick/003_cracker_box.h5
      - split: metadata
        path: pick/003_cracker_box.json
  - config_name: pick-004_sugar_box
    data_files:
      - split: trajectories
        path: pick/004_sugar_box.h5
      - split: metadata
        path: pick/004_sugar_box.json
  - config_name: pick-005_tomato_soup_can
    data_files:
      - split: trajectories
        path: pick/005_tomato_soup_can.h5
      - split: metadata
        path: pick/005_tomato_soup_can.json
  - config_name: pick-007_tuna_fish_can
    data_files:
      - split: trajectories
        path: pick/007_tuna_fish_can.h5
      - split: metadata
        path: pick/007_tuna_fish_can.json
  - config_name: pick-008_pudding_box
    data_files:
      - split: trajectories
        path: pick/008_pudding_box.h5
      - split: metadata
        path: pick/008_pudding_box.json
  - config_name: pick-009_gelatin_box
    data_files:
      - split: trajectories
        path: pick/009_gelatin_box.h5
      - split: metadata
        path: pick/009_gelatin_box.json
  - config_name: pick-010_potted_meat_can
    data_files:
      - split: trajectories
        path: pick/010_potted_meat_can.h5
      - split: metadata
        path: pick/010_potted_meat_can.json
  - config_name: pick-024_bowl
    data_files:
      - split: trajectories
        path: pick/024_bowl.h5
      - split: metadata
        path: pick/024_bowl.json
  - config_name: place-002_master_chef_can
    data_files:
      - split: trajectories
        path: place/002_master_chef_can.h5
      - split: metadata
        path: place/002_master_chef_can.json
  - config_name: place-003_cracker_box
    data_files:
      - split: trajectories
        path: place/003_cracker_box.h5
      - split: metadata
        path: place/003_cracker_box.json
  - config_name: place-004_sugar_box
    data_files:
      - split: trajectories
        path: place/004_sugar_box.h5
      - split: metadata
        path: place/004_sugar_box.json
  - config_name: place-005_tomato_soup_can
    data_files:
      - split: trajectories
        path: place/005_tomato_soup_can.h5
      - split: metadata
        path: place/005_tomato_soup_can.json
  - config_name: place-007_tuna_fish_can
    data_files:
      - split: trajectories
        path: place/007_tuna_fish_can.h5
      - split: metadata
        path: place/007_tuna_fish_can.json
  - config_name: place-008_pudding_box
    data_files:
      - split: trajectories
        path: place/008_pudding_box.h5
      - split: metadata
        path: place/008_pudding_box.json
  - config_name: place-009_gelatin_box
    data_files:
      - split: trajectories
        path: place/009_gelatin_box.h5
      - split: metadata
        path: place/009_gelatin_box.json
  - config_name: place-010_potted_meat_can
    data_files:
      - split: trajectories
        path: place/010_potted_meat_can.h5
      - split: metadata
        path: place/010_potted_meat_can.json
  - config_name: place-024_bowl
    data_files:
      - split: trajectories
        path: place/024_bowl.h5
      - split: metadata
        path: place/024_bowl.json

ManiSkill-HAB TidyHouse Dataset

Paper | Website | Code | Models | (Full) Dataset | Supplementary

Whole-body, low-level control/manipulation demonstration dataset for ManiSkill-HAB TidyHouse.

Dataset Details

Dataset Description

Demonstration dataset for ManiSkill-HAB TidyHouse. Each subtask/object combination (e.g pick 002_master_chef_can) has 1000 successful episodes (200 samples/demonstration) gathered using RL policies fitered for safe robot behavior with a rule-based event labeling system.

TidyHouse contains the Pick and Place subtasks. Relative to the other MS-HAB long-horizon tasks (PrepareGroceries, SetTable), TidyHouse Pick is approximately medium difficulty, while TidyHouse Place is medium-to-hard difficulty (on a scale of easy-medium-hard).

Related Datasets

Full information about the MS-HAB datasets (size, difficulty, links, etc), including the other long horizon tasks, are available on the ManiSkill-HAB website.

Uses

Direct Use

This dataset can be used to train vision-based learning from demonstrations and imitation learning methods, which can be evaluated with the MS-HAB environments. This dataset may be useful as synthetic data for computer vision tasks as well.

Out-of-Scope Use

While blind state-based policies can be trained on this dataset, it is recommended to train vision-based policies to handle collisions and obstructions.

Dataset Structure

Each subtask/object combination has files [SUBTASK]/[OBJECT].json and [SUBTASK]/[OBJECT].h5. The JSON file contains episode metadata, event labels, etc, while the HDF5 file contains the demonstration data.

Dataset Creation

The data is gathered using RL policies fitered for safe robot behavior with a rule-based event labeling system.

Bias, Risks, and Limitations

The dataset is purely synthetic.

While MS-HAB supports high-quality ray-traced rendering, this dataset uses ManiSkill's default rendering for data generation due to efficiency. However, users can generate their own data with the data generation code.

Citation

@article{shukla2024maniskillhab,
    author		 = {Arth Shukla and Stone Tao and Hao Su},
    title        = {ManiSkill-HAB: A Benchmark for Low-Level Manipulation in Home Rearrangement Tasks},
    journal      = {CoRR},
    volume       = {abs/2412.13211},
    year         = {2024},
    url          = {https://doi.org/10.48550/arXiv.2412.13211},
    doi          = {10.48550/ARXIV.2412.13211},
    eprinttype   = {arXiv},
    eprint       = {2412.13211},
    timestamp    = {Mon, 09 Dec 2024 01:29:24 +0100},
    biburl       = {https://dblp.org/rec/journals/corr/abs-2412-13211.bib},
    bibsource    = {dblp computer science bibliography, https://dblp.org}
}