---
license: mit
task_categories:
- robotics
---

# Dataset of Reactive Diffusion Policy

## Contents

- [Description](#description)
- [Structure](#structure)
- [Usage](#usage)

## Description

This is the raw dataset used in the paper
[Reactive Diffusion Policy: Slow-Fast Visual-Tactile Policy Learning for Contact-Rich Manipulation](https://huggingface.co/papers/2503.02881).

- [Paper](https://arxiv.org/pdf/2503.02881)
- [Project Homepage](https://reactive-diffusion-policy.github.io)
- [GitHub Repository](https://github.com/xiaoxiaoxh/reactive_diffusion_policy)

## Structure

We offer two versions of the dataset:
one is the [full dataset](https://huggingface.co/datasets/WendiChen/reactive_diffusion_policy_dataset/tree/main/dataset_full) used to train the models in our paper,
and the other is a [mini dataset](https://huggingface.co/datasets/WendiChen/reactive_diffusion_policy_dataset/tree/main/dataset_mini) for easier examination.
Both versions include subsets of peeling, wiping and lifting.
Each subset is structured as follows:

```
subset_name
├── seq_01.pkl
├── seq_02.pkl
├── ...
```
Note that we split the full lifting subset into 2 parts due to file size restrictions.

## Usage

Follow the [README in our GitHub repo](https://github.com/xiaoxiaoxh/reactive_diffusion_policy)
to [postprocess the data](https://github.com/xiaoxiaoxh/reactive_diffusion_policy#data-postprocessing)
and [train the model](https://github.com/xiaoxiaoxh/reactive_diffusion_policy#-training).