Datasets:

FreDERicia
/

SWAM-Pretrain

@@ -1,94 +1,121 @@
----
-pretty_name: SWAM Pretraining Data
-tags:
-  - robotics
-  - world-action-models
-  - video
-  - egocentric-video
-  - lerobot
-  - av1
----
-# SWAM-Pretrain
-**Scaling World Action Models with Cross-Modal Alignment**
-A collection of robot demonstrations, egocentric videos, and general videos prepared for world action model research. This repository distributes processed data from multiple upstream datasets in a consistent video format while preserving source-specific episode records and annotations.
-## Release status
-**This is an incremental release.** As of November 9, 2026, all **22 Galaxea archives (111.90 GB)** have been uploaded and checked against their local SHA-256 hashes. Anonymous download and extraction of its metadata and one complete chunk also passed validation, including Parquet loading and AV1 decoding for three sampled episodes. Further subsets are queued for batch upload; the complete collection is not yet available.
-The planned release contains **26 source directories, 1,238 TAR archives, and approximately 2.94 TB** (decimal units). It covers AgiBot World, RoboCoin, RoboMIND, InternData A1, Galaxea, ABC-130K, EgoVerse, and OpenVid. Panda70M, Ego4D, EgoSuite, and UMI are excluded from this release.
-Use the [Files and versions](https://huggingface.co/datasets/FreDERicia/SWAM-Pretrain/tree/main/data) tab to check which archives are currently available. Directory names preserve the original organization; resolution labels in those names describe the source data, not the resized videos.
-## Data format
-- **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 those two archives into a separate sample directory using the same pattern. Metadata may describe the full source dataset, so a partial download requires a loader restricted to the downloaded episodes. For reproducible downloads, add `--revision <commit-hash>` to the download command.
-Use a video reader with **AV1 decoding support**, such as a recent FFmpeg build with dav1d. Encoding libraries such as SVT-AV1 are not required to read these videos. Dataset schemas and action conventions remain source-specific; consult each source's `meta/info.json` before loading.
-## 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.

+---
+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/)
+**Scaling World Action Models with Cross-Modal Alignment**
+![SWAM research overview](visual-previews/teaser.png)
+*Research overview of the full pretraining data pool, model, and experiments. The public release scope is listed below.*
+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.
+## Release status
+**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.
+## 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.*
+[View source composition at full resolution](visual-previews/data-source-composition.png) · [View distribution statistics at full resolution](visual-previews/pretraining-composition-duration.png)
+## Data excluded from this release
+**Panda70M, Ego4D, and EgoSuite** are excluded due to upstream licensing and redistribution conditions. This repository does not distribute their original videos or processed video archives. Researchers should obtain these datasets through their official channels and follow the applicable agreements.
+**UMI** data is reserved for a **future release by Noematrix**. Its scope, timing, and access arrangements will follow Noematrix's official announcements.
+## Data format
+- **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.