README.md

---
pretty_name: Worldscape-MoE
license: cc-by-4.0
task_categories:
  - text-to-video
language:
  - en
size_categories:
  - 10K<n<100K
tags:
  - robotics
  - world-model
  - camera-control
  - mixture-of-experts
  - video
viewer: false
---

# Worldscape-MoE Dataset

This repository provides a 20,000-sample subset for unified
embodied world-model training: 5,000 samples each for camera control, dual-arm
control, dense hand/action-map control, and LIBERO single-arm control.

> [!IMPORTANT]
> This release is a subset of the full Worldscape-MoE training collection.
> 
- Project: https://worldscape-moe.com/
- Code: https://github.com/EmbodiedCity/Worldscape-MoE.code
- Model: https://huggingface.co/EmbodiedCity/Worldscape-MoE
- Paper: https://arxiv.org/abs/2607.03964

## Download

```bash
hf download EmbodiedCity/Worldscape-MoE-Dataset \
  --repo-type dataset \
  --local-dir datasets/Worldscape-MoE-Dataset
```

The four modality directories are distributed as compressed archives. Extract
them from the dataset root before training or inference:

```bash
cd datasets/Worldscape-MoE-Dataset
for archive in archives/*.tar.zst; do
  tar --zstd -xf "$archive"
done
```

This restores `data/camera`, `data/arm`, `data/action_map`, and `data/libero`
without changing the paths referenced by the metadata files.

## Contents

```text
.
├── README.md
├── LICENSE
├── DATASET_SOURCES.md
├── dataset_info.json
├── metadata/
│   ├── camera.json
│   ├── arm.json
│   ├── action_map.json
│   ├── libero.json
│   ├── train_3modal.json
│   └── train_4modal_libero.json
├── archives/
│   ├── worldscape-moe-camera.tar.zst
│   ├── worldscape-moe-arm.tar.zst
│   ├── worldscape-moe-action-map.tar.zst
│   └── worldscape-moe-libero.tar.zst
├── stats/{dual_arm_action_stats.json,libero_action_stats.json}
└── config/wan_civitai_5b.yaml
```

After extraction, the dataset also contains
`data/{camera,arm,action_map,libero}/{media,controls}/`. All paths are relative
to the dataset root. `train_3modal.json` contains 15,000 cases and
`train_4modal_libero.json` contains all 20,000 cases.

The released camera subset contains 5,000 RealEstate10K (RE10K) samples paired
with processed camera trajectories.

LIBERO cases are deterministic 17-frame windows. Multiple windows can share an
episode video, so physical media files are deduplicated.

## Metadata schema

Every row contains `type`, `file_path`, `text`, and `control_type`.

- Camera rows use `control_file_path` for camera poses.
- Action-map rows use `action_map_path` for the dense control video.
- Dual-arm and LIBERO rows use `ann_file` and `arm_action_key`.
- Windowed rows can include `start_frame`, `window_size`,
  `video_sample_stride`, and `video_sample_n_frames`.

```json
{
  "type": "video",
  "file_path": "data/libero/media/example.mp4",
  "ann_file": "data/libero/controls/example.json",
  "text": "put the object in the drawer",
  "control_type": "libero",
  "arm_action_key": "state",
  "start_frame": 12,
  "window_size": 17
}
```

Dual-arm annotations contain 14D `joint_action` values. LIBERO annotations
contain 7D `state` values. Their percentile statistics are stored separately.
LIBERO values are normalized before padding to the model's 14D action input.

## Use with Worldscape-MoE

```bash
DATA_ROOT=datasets/OpenSource_MOE \
bash scripts/wan2.2_fun/train_worldscape_moe_4modal_libero_5b_8gpu.sh
```

The manifests are training-oriented JSON arrays and the Dataset Viewer is
disabled. Load them directly with the Worldscape-MoE data loader.

## Reproducibility and validation

The release uses deterministic sampling with seed 42. Exact sample,
unique-file, and byte counts are recorded in `dataset_info.json`.

```bash
python tools/validate_opensource_dataset.py \
  datasets/OpenSource_MOE \
  --decode-samples-per-modality 8
```

The validator checks manifest membership, paths, action values and dimensions,
camera poses, percentile statistics, action windows, and sampled video
decoding.

## Sources and license

Collection-level provenance and attribution are listed in
`DATASET_SOURCES.md`. The dataset is released under the Creative Commons
Attribution 4.0 International license; see `LICENSE`.

When redistributing or adapting the dataset, cite Worldscape-MoE, retain this
dataset card, and preserve the upstream attributions listed in
`DATASET_SOURCES.md`.