README.md

---
pretty_name: "EgoSuite-Open10K · EgoDemo"
language:
- en
- zh
tags:
- robotics
- embodied-ai
- physical-ai
- egocentric-video
- human-demonstration
- multimodal
- demo
- streaming
viewer: false
configs:
- config_name: EgoDemo
  data_files:
  - split: train
    path: "manifests/EgoDemo/*.parquet"
---

<div align="center">

<img src="./assets/hero.png" width="100%" alt="EgoSuite-Open10K — real-world egocentric human data by Lightwheel" />

# EgoDemo

### An open, ungated trial pack sampled from the formal EgoSuite-Open10K repositories.

**20–50 selected hours · duplicate samples · full source provenance · excluded from the 10,000-hour total**

[Collection](https://huggingface.co/collections/xuejf/egosuite-open10k) · [EgoStandard](https://huggingface.co/datasets/xuejf/EgoStandard) · [EgoPro](https://huggingface.co/datasets/xuejf/EgoPro) · [中文](./README_zh.md) · [EgoSuite](https://lightwheel.ai/egosuite)

</div>

> **Public ungated landing card.** EgoDemo is publicly discoverable and remains ungated. The Dataset Viewer remains disabled until the audited demo manifests, media, approved license, and final citation are uploaded. Each published row must preserve source provenance and `counts_toward_10k=false`; public visibility does not imply that the final demo payload is complete.

## Start here

EgoDemo is the zero-friction entry point to EgoSuite-Open10K. It contains a curated 20–50-hour subset copied from the two formal repositories so users can inspect schemas, test loading and evaluate research fit before requesting gated access.

<img src="./assets/demo-provenance.svg" width="100%" alt="EgoDemo source provenance and counting rule" />

> **Counting rule:** every Demo row is a duplicate of a formal source row and carries `counts_toward_10k=false`. EgoDemo is never added to the official 10,000-hour total.

## The complete release

<img src="./assets/repository-map.svg" width="100%" alt="EgoSuite-Open10K collection and repository architecture" />

| Repository | Role | Configs | Formal hours | Access at release |
|---|---|---|---:|---|
| [EgoStandard](https://huggingface.co/datasets/xuejf/EgoStandard) | Head-view series | EgoStandard / EgoStand-motion / EgoFull | **9,000** | Manual gated |
| [EgoPro](https://huggingface.co/datasets/xuejf/EgoPro) | Head + wrist series | EgoPro / EgoPro-motion / EgoProMax | **1,000** | Manual gated |
| **EgoDemo** | Open duplicate trial pack | EgoDemo | 20–50, excluded from total | **Public, ungated** |

## Load the trial pack

```python
from datasets import load_dataset

demo = load_dataset(
    "LightwheelAI/EgoDemo",
    "EgoDemo",
    split="train",
    streaming=True,
)

first = next(iter(demo))
print(first["source_repo"], first["source_config"], first["source_clip_id"])
```

## Provenance contract

| Field | Type | Meaning |
|---|---|---|
| `demo_clip_id` | string | Demo-local anonymous row key |
| `source_repo` | string | `LightwheelAI/EgoStandard` or `LightwheelAI/EgoPro` |
| `source_config` | string | One of the six formal SKU names |
| `source_clip_id` | string | Stable clip key in the source release |
| `source_revision` | string | Immutable source commit or release tag |
| `counts_toward_10k` | bool | Must always be `false` |
| `scene_family` | string | One of seven top-level scene families |
| `duration_s` | float32 | Audited usable duration in seconds |
| `head_video_ref` | string | Demo-local or immutable source media reference |
| `wrist_video_ref` | string, nullable | Present for sampled Pro-series rows |
| `hand_pose_3d_ref` | string | Hand-pose reference |
| `body_pose_3d_ref` | string, nullable | Body-pose reference when available |
| `semantics_v7_ref` | string, nullable | V7 semantic reference when available |
| `sha256` | string | Integrity checksum |

The schema is defined in [`metadata/schema.json`](./metadata/schema.json). The final `statistics.json` must report the exact Demo duration and source-config mix.

## Repository layout

```text
EgoDemo/
├── README.md
├── README_zh.md
├── assets/
├── manifests/EgoDemo/part-*.parquet
├── media/<selected-demo-shards>.tar
├── annotations/<selected-demo-shards>.tar
└── metadata/
    ├── source_map.parquet
    ├── schema.json
    ├── statistics.json
    └── checksums.sha256
```

## Scene coverage

<img src="./assets/seven-scenarios.png" width="100%" alt="Seven scene families covered by EgoSuite-Open10K" />

The Demo should sample across **Home, Hospitality, Retail, Sports, Logistics, Office, and Industry**, while keeping the precise mix traceable to the formal source rows. Selection criteria and exact hours must be disclosed in the release manifest.

## Access and responsible use

EgoDemo is **public and ungated**, but open access does not remove the dataset license, participant-consent scope, privacy requirements, or downstream-use restrictions.

Users must not attempt to identify participants, reconstruct sensitive locations, or use the data for surveillance, profiling, or harmful applications.

## License, citation, and contact

The approved dataset license and citation author list are still pending. Public visibility and ungated access do not grant rights beyond the governing terms and provenance rules approved for the demo source data.

For product information, visit [Lightwheel EgoSuite](https://lightwheel.ai/egosuite). For release or collaboration inquiries, use the official [Lightwheel contact form](https://lightwheel.ai/contact).

```bibtex
@dataset{lightwheel_egosuite_open10k_egodemo_2026,
  author    = {{Lightwheel}},
  title     = {EgoSuite-Open10K: EgoDemo},
  year      = {2026},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/xuejf/EgoDemo}
}
```

> Citation metadata is a release-candidate template until the approved author list and publication record are locked.

---
pretty_name: "EgoSuite-Open10K · EgoDemo"
language:
- en
- zh
tags:
- robotics
- embodied-ai
- physical-ai
- egocentric-video
- human-demonstration
- multimodal
- demo
- streaming
viewer: false
configs:
- config_name: EgoDemo
  data_files:
  - split: train
    path: "manifests/EgoDemo/*.parquet"
---

<div align="center">

<img src="./assets/hero.png" width="100%" alt="EgoSuite-Open10K — real-world egocentric human data by Lightwheel" />

# EgoDemo

### An open, ungated trial pack sampled from the formal EgoSuite-Open10K repositories.

**20–50 selected hours · duplicate samples · full source provenance · excluded from the 10,000-hour total**

[Collection](https://huggingface.co/collections/xuejf/egosuite-open10k) · **Private staging · xuejf namespace** · [EgoStandard](https://huggingface.co/datasets/xuejf/EgoStandard) · [EgoPro](https://huggingface.co/datasets/xuejf/EgoPro) · [中文](./README_zh.md) · [EgoSuite](https://lightwheel.ai/egosuite)

</div>

> **Private staging preview.** This repository currently contains the Dataset Card, visual assets, manifest contract, and schema only. The audited Parquet manifests, media shards, approved license, and final citation have not been uploaded; no access rights to the unpublished 10K corpus are granted by this preview.

## Start here

EgoDemo is the zero-friction entry point to EgoSuite-Open10K. It contains a curated 20–50-hour subset copied from the two formal repositories so users can inspect schemas, test loading and evaluate research fit before requesting gated access.

<img src="./assets/demo-provenance.svg" width="100%" alt="EgoDemo source provenance and counting rule" />

> **Counting rule:** every Demo row is a duplicate of a formal source row and carries `counts_toward_10k=false`. EgoDemo is never added to the official 10,000-hour total.

## The complete release

<img src="./assets/four-repository-map.png" width="100%" alt="EgoSuite-Open10K collection and repository architecture" />

| Repository | Role | Configs | Formal hours | Access at release |
|---|---|---|---:|---|
| [EgoStandard](https://huggingface.co/datasets/xuejf/EgoStandard) | Head-view series | EgoStandard / EgoStand-motion / EgoFull | **9,000** | Manual gated |
| [EgoPro](https://huggingface.co/datasets/xuejf/EgoPro) | Head + wrist series | EgoPro / EgoPro-motion / EgoProMax | **1,000** | Manual gated |
| **EgoDemo** | Open duplicate trial pack | EgoDemo | 20–50, excluded from total | **Public, ungated** |

## Load the trial pack

```python
from datasets import load_dataset

demo = load_dataset(
    "xuejf/EgoDemo",
    "EgoDemo",
    split="train",
    streaming=True,
)

first = next(iter(demo))
print(first["source_repo"], first["source_config"], first["source_clip_id"])
```

## Provenance contract

| Field | Type | Meaning |
|---|---|---|
| `demo_clip_id` | string | Demo-local anonymous row key |
| `source_repo` | string | `xuejf/EgoStandard` or `xuejf/EgoPro` |
| `source_config` | string | One of the six formal SKU names |
| `source_clip_id` | string | Stable clip key in the source release |
| `source_revision` | string | Immutable source commit or release tag |
| `counts_toward_10k` | bool | Must always be `false` |
| `scene_family` | string | One of seven top-level scene families |
| `duration_s` | float32 | Audited usable duration in seconds |
| `head_video_ref` | string | Demo-local or immutable source media reference |
| `wrist_video_ref` | string, nullable | Present for sampled Pro-series rows |
| `hand_pose_3d_ref` | string | Hand-pose reference |
| `body_pose_3d_ref` | string, nullable | Body-pose reference when available |
| `semantics_v7_ref` | string, nullable | V7 semantic reference when available |
| `sha256` | string | Integrity checksum |

The schema is defined in [`metadata/schema.json`](./metadata/schema.json). The final `statistics.json` must report the exact Demo duration and source-config mix.

## Repository layout

```text
EgoDemo/
├── README.md
├── README_zh.md
├── assets/
├── manifests/EgoDemo/part-*.parquet
├── media/<selected-demo-shards>.tar
├── annotations/<selected-demo-shards>.tar
└── metadata/
    ├── source_map.parquet
    ├── schema.json
    ├── statistics.json
    └── checksums.sha256
```

## Scene coverage

<img src="./assets/trial-pack.png" width="100%" alt="EgoDemo trial pack contents" />

<img src="./assets/seven-environments.png" width="100%" alt="Seven scene families covered by EgoSuite-Open10K" />

The Demo should sample across **Home, Hospitality, Retail, Sports, Logistics, Office, and Industry**, while keeping the precise mix traceable to the formal source rows. Selection criteria and exact hours must be disclosed in the release manifest.

## Access and responsible use

EgoDemo is intended to be **public and ungated**, but open access does not remove the dataset license, participant-consent scope, privacy requirements, or downstream-use restrictions.

Users must not attempt to identify participants, reconstruct sensitive locations, or use the data for surveillance, profiling, or harmful applications.

## License, citation, and contact

The approved dataset license and citation author list must be added before public release. EgoDemo should use the same governing terms and provenance rules as its formal source data unless Lightwheel legal explicitly approves a separate demo license.

For product information, visit [Lightwheel EgoSuite](https://lightwheel.ai/egosuite). For release or collaboration inquiries, use the official [Lightwheel contact form](https://lightwheel.ai/contact).

```bibtex
@dataset{lightwheel_egosuite_open10k_egodemo_2026,
  author    = {{Lightwheel}},
  title     = {EgoSuite-Open10K: EgoDemo},
  year      = {2026},
  publisher = {Hugging Face},
  url       = {https://huggingface.co/datasets/xuejf/EgoDemo}
}
```

> Citation metadata is a release-candidate template until the approved author list and publication record are locked.