FastWAM Checkpoints

This repository releases the official FastWAM checkpoints for the open-source release of Fast-WAM: Do World Action Models Need Test-time Future Imagination?

This Hugging Face model repository contains:

README.md
libero_uncond_2cam224.pt
libero_uncond_2cam224_dataset_stats.json
robotwin_uncond_3cam_384.pt
robotwin_uncond_3cam_384_dataset_stats.json

Summary

• Project: FastWAM
• Framework: PyTorch
• Checkpoint count: 2
• Checkpoint format: .pt
• Supported benchmarks in this repo: LIBERO, RoboTwin

Project

• Project page: https://yuantianyuan01.github.io/FastWAM/
• Paper: https://arxiv.org/abs/2603.16666
• Code repository: https://github.com/yuantianyuan01/FastWAM

Included checkpoints

• libero_uncond_2cam224.pt: checkpoint for task=libero_uncond_2cam224_1e-4
• libero_uncond_2cam224_dataset_stats.json: dataset stats for LIBERO evaluation
• robotwin_uncond_3cam_384.pt: checkpoint for task=robotwin_uncond_3cam_384_1e-4
• robotwin_uncond_3cam_384_dataset_stats.json: dataset stats for RoboTwin evaluation

Citation

If you use these checkpoints, please cite the Fast-WAM paper.

@misc{yuan2026fastwam,
  title={Fast-WAM: Do World Action Models Need Test-time Future Imagination?},
  author={Tianyuan Yuan and Zibin Dong and Yicheng Liu and Hang Zhao},
  year={2026},
  note={arXiv preprint arXiv:2603.16666}
}