[docs] update dataset card

README.md

@@ -1,3 +1,43 @@
----
-license: mit
----
+---
+license: mit
+task_categories:
+- robotics
+---
+
+# Dataset of ImplicitRDP
+
+## Contents
+
+- [Description](#description)
+- [Structure](#structure)
+- [Usage](#usage)
+
+## Description
+
+This is the raw dataset used in the paper
+[ImplicitRDP: An End-to-End Visual-Force Diffusion Policy with Structural Slow-Fast Learning](https://implicit-rdp.github.io).
+
+- [Paper](https://arxiv.org/pdf/2512.10946)
+- [Project Homepage](https://implicit-rdp.github.io)
+- [GitHub Repository](https://github.com/Chen-Wendi/ImplicitRDP)
+- [Pretrained Models](https://huggingface.co/WendiChen/ImplicitRDP_model)
+
+## Structure
+
+We offer two versions of the dataset:
+one is the [full dataset](https://huggingface.co/datasets/WendiChen/ImplicitRDP_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/ImplicitRDP_dataset/tree/main/dataset_mini) for easier examination.
+Both versions include the raw subsets of box flipping and switch toggling.
+
+Each subset is structured as follows:
+```
+task_name
+├── 1.pkl
+├── 2.pkl
+├── ...
+```
+
+## Usage
+
+Follow the [README in our GitHub repo](https://github.com/Chen-Wendi/ImplicitRDP)
+to [train the model](https://github.com/Chen-Wendi/ImplicitRDP#-training).