KAI0

TODO

• The advantage label will be coming soon.

Contents

• About the Dataset
• Load the Dataset
• Download the Dataset
• Dataset Structure
  • Folder hierarchy
  • Details
• License and Citation

About the Dataset

• ~134 hours real world scenarios

• Main Tasks

  • FlattenFold
    • Single task
    • Initial state: T-shirts are randomly tossed onto the table, presenting random crumpled configurations
    • Manipulation task: Operate the robotic arm to unfold the garment, then fold it
  • HangCloth
    • Single task
    • Initial state: Hanger is randomly placed, garment is randomly positioned on the table
    • Manipulation task: Operate the robotic arm to thread the hanger through the garment, then hang it on the rod
  • TeeShirtSort
    • Garment classification and arrangement task
    • Initial state: Randomly pick a garment from the laundry basket
    • Classification: Determine whether the garment is a T-shirt or a dress shirt
    • Manipulation task:
      • If it is a T-shirt, fold the garment
      • If it is a dress shirt, expose the collar, then push it to one side of the table

• Count of the dataset

  Task	Base (episodes count/hours)	DAgger (episodes count/hours)	Total(episodes count/hours)
  FlattenFold	3,055/~42 hours	3,457/ ~13 Hours	6,512 /~55 hours
  HangCloth	6954/~61 hours	686/~12 hours	7640/~73 hours
  TeeShirtSort	5988/~31 hours	769/~22 hours	6757/~53 hours
  Total	15,997/~134 hours	4,912/~47 hours	20,909/~181 hours

Load the dataset

• This dataset was created using LeRobot
• The dataset's version is LeRobotDataset v2.1

For LeRobot version < 0.4.0

Choose the appropriate import based on your version:

Version	Import Path
<= 0.1.0	from lerobot.common.datasets.lerobot_dataset import LeRobotDataset
> 0.1.0 and < 0.4.0	from lerobot.datasets.lerobot_dataset import LeRobotDataset

# For version <= 0.1.0
from lerobot.common.datasets.lerobot_dataset import LeRobotDataset

# For version > 0.1.0 and < 0.4.0
from lerobot.datasets.lerobot_dataset import LeRobotDataset

# Load the dataset
dataset = LeRobotDataset(repo_id='where/the/dataset/you/stored')

For LeRobot version >= 0.4.0

You need to migrate the dataset from v2.1 to v3.0 first. See the official documentation: Migrate the dataset from v2.1 to v3.0

python -m lerobot.datasets.v30.convert_dataset_v21_to_v30 --repo-id=<HF_USER/DATASET_ID>

Download the Dataset

Python Script

from huggingface_hub import hf_hub_download, snapshot_download
from datasets import load_dataset

# Download a single file
hf_hub_download(
    repo_id="OpenDriveLab-org/kai0", 
    filename="episodes.jsonl",
    subfolder="meta",
    repo_type="dataset",
    local_dir="where/you/want/to/save"
)

# Download a specific folder
snapshot_download(
    repo_id="OpenDriveLab-org/kai0", 
    local_dir="/where/you/want/to/save",
    repo_type="dataset",
    allow_patterns=["data/*"]
)

# Load the entire dataset
dataset = load_dataset("OpenDriveLab-org/kai0") 

Terminal (CLI)

# Download a single file
hf download OpenDriveLab-org/kai0 \
    --include "meta/info.json" \
    --repo-type dataset \
    --local-dir "/where/you/want/to/save"

# Download a specific folder
hf download OpenDriveLab-org/kai0 \
    --repo-type dataset \
    --include "meta/*" \
    --local-dir "/where/you/want/to/save"

# Download the entire dataset
hf download OpenDriveLab-org/kai0 \
    --repo-type dataset \
    --local-dir "/where/you/want/to/save"

Dataset Structure

Folder hierarchy

Under each task directory, data is partitioned into two subsets: base and dagger.

• base contains original demonstration trajectories of robotic arm manipulation for garment arrangement tasks.
• dagger
  contains on-policy recovery trajectories collected via iterative DAgger, designed to populate failure recovery modes absent in static demonstrations.

Kai0-data/
├── FlattenFold/
│   ├── base/
│   │   ├── data/
│   │   │   ├── chunk-000/
│   │   │   │   ├── episode_000000.parquet
│   │   │   │   ├── episode_000001.parquet
│   │   │   │   └── ...
│   │   │   └── ...
│   │   ├── videos/
│   │   │   ├── chunk-000/
│   │   │   │   ├── observation.images.hand_left/
│   │   │   │   │   ├── episode_000000.mp4
│   │   │   │   │   ├── episode_000001.mp4
│   │   │   │   │   └── ...
│   │   │   │   ├── observation.images.hand_right/
│   │   │   │   │   ├── episode_000000.mp4
│   │   │   │   │   ├── episode_000001.mp4
│   │   │   │   │   └── ...
│   │   │   │   ├── observation.images.top_head/
│   │   │   │   │   ├── episode_000000.mp4
│   │   │   │   │   ├── episode_000001.mp4
│   │   │   │   │   └── ...
│   │   │   │   └── ...
│   │   │   └── ...
│   │   └── meta/
│   │       ├── info.json
│   │       ├── episodes.jsonl
│   │       ├── tasks.jsonl
│   │       └── episodes_stats.jsonl
│   └── dagger/
├── HangCloth/
│   ├── base/
│   └── dagger/
├── TeeShirtSort/
│   ├── base/
│   └── dagger/
└── README.md

Details

info.json

the basic struct of the info.json

{
    "codebase_version": "v2.1",
    "robot_type": "agilex",
    "total_episodes": ...,  # the total episodes in the dataset
    "total_frames": ...,    # The total number of video frames in any single camera perspective
    "total_tasks": ...,     # Total number of tasks
    "total_videos": ...,    # The total number of videos from all camera perspectives in the dataset
    "total_chunks": ...,    # The number of chunks in the dataset
    "chunks_size": ...,     # The max number of episodes in a chunk
    "fps": ...,             # Video frame rate per second
    "splits": {             # how to split the dataset
        "train": ...       
    },
    "data_path": "data/chunk-{episode_chunk:03d}/episode_{episode_index:06d}.parquet",
    "video_path": "videos/chunk-{episode_chunk:03d}/{video_key}/episode_{episode_index:06d}.mp4",
    "features": {
        "observation.images.top_head": {   # the camera perspective
            "dtype": "video",
            "shape": [
                480,
                640,
                3
            ],
            "names": [
                "height",
                "width",
                "channel"
            ],
            "info": {
                "video.height": 480,
                "video.width": 640,
                "video.codec": "av1",
                "video.pix_fmt": "yuv420p",
                "video.is_depth_map": false,
                "video.fps": 30,
                "video.channels": 3,
                "has_audio": false
            }
        },
        "observation.images.hand_left": {   # the camera perspective
            ...
        },
        "observation.images.hand_right": {   # the camera perspective
            ...
        },
        "observation.state": {
            "dtype": "float32",
            "shape": [
                14
            ],
            "names": null
        },
        "action": {
            "dtype": "float32",
            "shape": [
                14
            ],
            "names": null
        },
        "timestamp": {
            "dtype": "float32",
            "shape": [
                1
            ],
            "names": null
        },
        "frame_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "episode_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        },
        "task_index": {
            "dtype": "int64",
            "shape": [
                1
            ],
            "names": null
        }
    }
}

Parquet file format

Field Name	shape	Meaning
observation.state	[N, 14]	left [:, :6], right [:, 7:13], joint angle left[:, 6], right [:, 13] , gripper open range
action	[N, 14]	left [:, :6], right [:, 7:13], joint angle left[:, 6], right [:, 13] , gripper open range
timestamp	[N, 1]	Time elapsed since the start of the episode (in seconds)
frame_index	[N, 1]	Index of this frame within the current episode (0-indexed)
episode_index	[N, 1]	Index of the episode this frame belongs to
index	[N, 1]	Global unique index across all frames in the dataset
task_index	[N, 1]	Index identifying the task type being performed

tasks.jsonl

Contains task language prompts (natural language instructions) that specify the manipulation task to be performed. Each entry maps a task_index to its corresponding task description, which can be used for language-conditioned policy training.

License and Citation

All the data and code within this repo are under . Please consider citing our project if it helps your research.

@misc{,
  title={},
  author={},
  howpublished={\url{}},
  year={}
}