Use English README with SWAM Pretrain styling
Browse files
README.md
CHANGED
|
@@ -13,33 +13,37 @@ tags:
|
|
| 13 |
|
| 14 |
# SWAM-Posttrain
|
| 15 |
|
| 16 |
-
[](https://char-lotte-xia.github.io/swam-official/)
|
| 17 |
|
| 18 |
+
<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="https://huggingface.co/datasets/FreDERicia/SWAM-Pretrain/resolve/main/visual-previews/link-github.png" alt="GitHub code" width="108" height="35" 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="https://huggingface.co/datasets/FreDERicia/SWAM-Pretrain/resolve/main/visual-previews/link-website.png" alt="Project website" width="108" height="35" 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="https://huggingface.co/datasets/FreDERicia/SWAM-Pretrain/resolve/main/visual-previews/link-arxiv.png" alt="arXiv paper" width="108" height="35" style="display: inline-block; margin: 0; vertical-align: middle;"></a></div>
|
| 19 |
|
| 20 |
+
This repository provides the RoboTwin Clean posttraining data used in SWAM: Scaling World Action Models with Cross-Modal Alignment.
|
| 21 |
|
| 22 |
+
The release contains 50 tasks and 2,500 episodes, distributed as 50 TAR archives totaling approximately 3.94 GB. [Browse all data](https://huggingface.co/datasets/FreDERicia/SWAM-Posttrain/tree/main/data)
|
| 23 |
|
| 24 |
+
## Data format and directory structure
|
| 25 |
|
| 26 |
+
Each task follows a LeRobot-style layout with `data/`, `videos/`, and `meta/` directories, containing actions and states, videos, language instructions, and statistics. Original episode IDs are preserved. Latents are not included.
|
| 27 |
|
| 28 |
+
The archives retain 16 episodes listed in the per-task `meta/episode_exclusions_20261023.jsonl` files. Apply these exclusions during training to obtain 2,484 eligible episodes. Task instructions are stored in `meta/tasks_v1.jsonl`.
|
| 29 |
|
| 30 |
+
## Download and extract
|
| 31 |
|
| 32 |
+
Install the Hugging Face CLI and sign in with an account that has access to this private repository.
|
| 33 |
|
| 34 |
```bash
|
| 35 |
pip install -U huggingface_hub
|
| 36 |
hf auth login
|
| 37 |
+
```
|
| 38 |
+
|
| 39 |
+
Download the dataset.
|
| 40 |
|
| 41 |
+
```bash
|
| 42 |
hf download FreDERicia/SWAM-Posttrain \
|
| 43 |
--repo-type dataset --local-dir ./SWAM-Posttrain
|
| 44 |
```
|
| 45 |
|
| 46 |
+
Extract all task archives into the same directory on Linux or macOS.
|
| 47 |
|
| 48 |
```bash
|
| 49 |
mkdir -p ./lerobot_robotwin_eef_clean_50
|
|
|
|
| 52 |
done
|
| 53 |
```
|
| 54 |
|
| 55 |
+
The extracted layout is `lerobot_robotwin_eef_clean_50/<task-directory>/{data,videos,meta}/`.
|
| 56 |
+
|
| 57 |
+
## Source terms and attribution
|
| 58 |
+
|
| 59 |
+
The source data declares a [CC BY-NC-SA 4.0](https://creativecommons.org/licenses/by-nc-sa/4.0/) license. Please follow the upstream license and attribution requirements. The original documentation is retained in [SOURCE_README.md](https://huggingface.co/datasets/FreDERicia/SWAM-Posttrain/blob/main/SOURCE_README.md).
|
| 60 |
|
| 61 |
+
## Related resources
|
| 62 |
|
| 63 |
+
[Model checkpoints](https://huggingface.co/FreDERicia/SWAM) · [Pretraining data](https://huggingface.co/datasets/FreDERicia/SWAM-Pretrain)
|
| 64 |
|
| 65 |
+
## Code and citation
|
| 66 |
|
| 67 |
+
Code is available in the [SWAM code repository](https://github.com/shaaaaron/SWAM).
|
| 68 |
|
| 69 |
+
If you use this dataset, please cite the SWAM paper.
|
| 70 |
|
| 71 |
```bibtex
|
| 72 |
@misc{zhou2026swam,
|