---
pretty_name: SWAM Pretraining Data
tags:
  - robotics
  - world-action-models
  - video
  - egocentric-video
  - lerobot
  - av1
---

# SWAM-Pretrain

[![SWAM project cover](visual-previews/swam-cover-cropped.png)](https://char-lotte-xia.github.io/swam-official/)

<div style="display: flex; justify-content: center; align-items: center; gap: 8px; flex-wrap: wrap; margin: 12px 0 20px;"><a style="display: inline-flex; margin: 0; line-height: 0;" href="https://github.com/shaaaaron/SWAM"><img src="visual-previews/link-github.png" alt="GitHub code" width="124" height="36" style="display: inline-block; margin: 0; vertical-align: middle;"></a><a style="display: inline-flex; margin: 0; line-height: 0;" href="https://char-lotte-xia.github.io/swam-official/"><img src="visual-previews/link-website.png" alt="Project website" width="124" height="36" style="display: inline-block; margin: 0; vertical-align: middle;"></a><a style="display: inline-flex; margin: 0; line-height: 0;" href="https://arxiv.org/abs/2611.12870"><img src="visual-previews/link-arxiv.png" alt="arXiv paper" width="124" height="36" style="display: inline-block; margin: 0; vertical-align: middle;"></a></div>

SWAM pretraining data is organized into **General, Ego, UMI, and Robotics**, covering general videos, egocentric videos, UMI manipulation demonstrations, and robot demonstrations. We standardize the video format, preserve source-specific episode records and annotations, and add detailed annotations for manipulation data. This repository releases the publicly distributable portion in stages.

**Data is being uploaded in batches. See the [repository file list](https://huggingface.co/datasets/FreDERicia/SWAM-Pretrain/tree/main/data) for currently available downloads.**

## Data scope

| Category | Data sources | Release scope |
| --- | --- | --- |
| General | OpenVid, Panda70M | OpenVid included; Panda70M excluded from this release |
| Ego | EgoVerse, Ego4D, EgoSuite | EgoVerse included; Ego4D and EgoSuite excluded from this release |
| UMI | UMI manipulation demonstrations | Awaiting a future release by Noematrix |
| Robotics | AgiBot World, RoboCoin, RoboMIND, InternData A1, Galaxea, ABC-130K | Released in batches according to the release manifest |

The planned release covers **26 source directories, approximately 2.94 TB** (decimal units). Directory names preserve the original organization; resolution labels describe the source data specifications.

**Why some data is excluded:** Panda70M, Ego4D, and EgoSuite are subject to upstream licensing and redistribution conditions; please obtain them through their official channels. UMI data awaits a **future release by Noematrix**, with release arrangements to be announced through its official channels.

## Research data statistics

The figures below describe the paper's **full pretraining data pool of approximately 47,000 hours**, including Panda70M, Ego4D, EgoSuite, and UMI, which are excluded from this release. These figures do not represent the amount of data publicly released or uploaded to this repository.

![Research data statistics: source composition on the left; video hours, episode shares, and duration distributions on the right](visual-previews/data-statistics-side-by-side.png)

*Left: data sources and video hours. Right: shares of total video hours and episodes, and video-duration distributions within each category.*

## Data format and directory structure

- **Video:** 320 pixels wide by 256 pixels high, AV1 encoded in MP4 containers. Source frame rates and temporal ordering are preserved.
- **Episode records:** Parquet files retain the actions, states, and language annotations provided by each source. Available fields vary across datasets.
- **Metadata:** Source-specific metadata is retained, with video-related dimensions and codec information updated for the processed files.
- **Layout:** LeRobot-style `data/`, `meta/`, and `videos/` directories are packaged as TAR archives. This is an archive distribution, not a single uniform tabular dataset.

Each source directory contains `metadata.tar` and one or more `chunk-*.tar` archives. Extract all archives for a source into the **same source directory**. For example:

```text
archives/data/galaxea/
  metadata.tar
  chunk-000.tar
  ...
  chunk-020.tar

# After extraction:
data/galaxea/
  meta/
  data/chunk-000/...
  videos/chunk-000/...
  ...
```

## Download and extract

Install the Hugging Face CLI:

```bash
pip install -U huggingface_hub
```

Download the complete Galaxea subset:

```bash
hf download FreDERicia/SWAM-Pretrain \
  --repo-type dataset \
  --include "data/galaxea/*.tar" \
  --local-dir ./archives
```

Extract it on Linux or macOS:

```bash
mkdir -p ./data/galaxea
for archive in ./archives/data/galaxea/*.tar; do
  tar -xf "$archive" -C ./data/galaxea
done
```

For a smaller download, fetch `metadata.tar` and one chunk:

```bash
hf download FreDERicia/SWAM-Pretrain \
  data/galaxea/metadata.tar data/galaxea/chunk-019.tar \
  --repo-type dataset --local-dir ./sample-archives
```

Extract archives from the same source into the same directory. For partial downloads, load only the corresponding episodes. Video readers need **AV1 decoding support**; consult each source's `meta/info.json` for field definitions.

## Source terms and attribution

This collection includes data from multiple upstream projects. Their respective licenses, attribution requirements, and usage restrictions continue to apply, including non-commercial restrictions where specified. Processing and repackaging do not grant additional rights or place the entire collection under a single permissive license.

## Code

The accompanying [SWAM code repository](https://github.com/shaaaaron/SWAM) integrates RLinf, DreamZero, DMuon, and RoboTwin, with entry points for training, checkpoint conversion, and RoboTwin evaluation. The code repository is currently private and requires authorized access.