README.md

---
library_name: pytorch
pipeline_tag: robotics
tags:
- robotics
- world-action-model
- manipulation
---

# PrismWAM

Checkpoints for **PrismWAM: Decoupling State Representations in World Action Models**.

[Code](https://github.com/THUSI-Lab/PrismWAM) · [Project page](https://prism-wam.github.io/) · [arXiv](https://arxiv.org/abs/2610.26564)

| Folder | Benchmark | Actions | Result |
| --- | --- | --- | ---: |
| `robotwin/` | RoboTwin 2.0, Clean → Randomized | 14-D absolute joints | 70.57 |
| `robodojo/` | RoboDojo | 14-D absolute joints | 25.10 / 18.55 |
| `libero/` | LIBERO, 4 suites | 7-D delta end effector + gripper | 99.35 |
| `init/` | training initializers: depth VAEs, RGB VAE, action-expert init | | |

Success rates in %; RoboDojo is score / success rate. Each model folder holds `config.yaml`,
`policy.pt`, the RGB and depth VAEs, and its action statistics.

## Usage

Install the [code](https://github.com/THUSI-Lab/PrismWAM), then from the repository root:

```bash
hf download THUSI-Lab/PrismWAM --include "robotwin/*" --local-dir weights
```

```python
from prismwam import load_pretrained

policy = load_pretrained("robotwin")   # also fetches the T5 text encoder on first use
observation["prompt"] = policy.format_prompt("beat the block with the hammer")
out = policy.predict(observation, seed=0)        # out["action"]: 32 x 14
```

The observation format, benchmark evaluation and training recipes are documented in the code
repository.

## Citation

```bibtex
@misc{prism,
  title         = {PrismWAM: Decoupling State Representations in World Action Models},
  author        = {Xin, Tianyu and Zhong, Wanqi and Hu, Bowen and Guo, Zhangchi and Wang, Lijie and Wang, Yiming and Liao, Haoran and Yu, Heyang and Tian, Xiaoyu and Yu, Zhiding and Li, Yiming},
  year          = {2026},
  eprint        = {2610.26564},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CV},
  url           = {https://arxiv.org/abs/2610.26564}
}
```