---
pipeline_tag: robotics
library_name: diffusers
license: mit
---

# Real-Time Iteration Scheme for Diffusion Policy (RTI-DP)

This repository contains the official model weights and code for the paper: **"Real-Time Iteration Scheme for Diffusion Policy"**.

- 📚 [Paper](https://huggingface.co/papers/2508.05396)
- 🌐 [Project Page](https://rti-dp.github.io/)
- 💻 [Code](https://github.com/RTI-DP/rti-dp)

RTI-DP enables fast inference in diffusion-based robotic policies by initializing each denoising step with the previous prediction — no retraining, no distillation.

<div align="center">
  <img src="https://rti-dp.github.io/static/images/rti.png" alt="RTI-DP Teaser" width="600"/>
</div>

## Usage

This model is designed to be used with its official codebase. For detailed installation instructions, environment setup, and further information, please refer to the [official GitHub repository](https://github.com/RTI-DP/rti-dp), which is based on [Diffusion Policy](https://github.com/real-stanford/diffusion_policy).

### Evaluation

To evaluate RTI-DP policies with DDPM, you can use the provided script from the repository:

```shell
python ../eval_rti.py --config-name=eval_diffusion_rti_lowdim_workspace.yaml
```

For RTI-DP-scale checkpoints, refer to the [duandaxia/rti-dp-scale](https://huggingface.co/duandaxia/rti-dp-scale) on Hugging Face.

## Citation

If you find our work useful, please consider citing our paper:

```bibtex
@misc{duan2025realtimeiterationschemediffusion,
      title={Real-Time Iteration Scheme for Diffusion Policy}, 
      author={Yufei Duan and Hang Yin and Danica Kragic},
      year={2025},
      eprint={2508.05396},
      archivePrefix={arXiv},
      primaryClass={cs.RO},
      url={https://arxiv.org/abs/2508.05396}, 
}
```

## Acknowledgements

We thank the authors of [Diffusion Policy](https://github.com/real-stanford/diffusion_policy), [Consistency Policy](https://github.com/Aaditya-Prasad/Consistency-Policy/) and [Streaming Diffusion Policy](https://github.com/Streaming-Diffusion-Policy/streaming_diffusion_policy/) for sharing their codebase.