Add robotics task category and improve dataset card

README.md

@@ -1,11 +1,39 @@
 ---
 license: apache-2.0
+task_categories:
+- robotics
 ---
 
 # RoboMME Training Data (TFDS Format)
 
-[Arxiv Paper](https://arxiv.org/abs/2603.04639) |  [HF Paper](https://huggingface.co/papers/2603.04639) |  [Website](https://robomme.github.io/) | [Benchmark Code](https://github.com/RoboMME/robomme_benchmark) | [Policy Learning Code](https://github.com/RoboMME/robomme_policy_learning)
+[Arxiv Paper](https://arxiv.org/abs/2603.04639) | [HF Paper](https://huggingface.co/papers/2603.04639) | [Website](https://robomme.github.io/) | [Benchmark Code](https://github.com/RoboMME/robomme_benchmark) | [Policy Learning Code](https://github.com/RoboMME/robomme_policy_learning)
 
-This repo contains tensorflow dataset format of RoboMME training data. It's mainly used for OpenVLA/MemoryVLA models.
+This repository contains the training data for **RoboMME**, a large-scale standardized benchmark for evaluating and advancing Vision-Language-Action (VLA) models in long-horizon, history-dependent robotic manipulation scenarios. 
 
-Our adaptation of MemoryVLA on RoboMME can be found [here](https://github.com/RoboMME/MemoryVLA).
+The data is provided in **TensorFlow Dataset (TFDS)** format and is primarily used for training models such as OpenVLA and MemoryVLA.
+
+## Benchmark Overview
+
+RoboMME evaluates 16 manipulation tasks across four categories of memory:
+
+| Suite      | Focus             | Task IDs                                                                 |
+| ---------- | ----------------- | --------------------------------------------------------------------- |
+| **Counting**   | Temporal memory   | BinFill, PickXtimes, SwingXtimes, StopCube                            |
+| **Permanence** | Spatial memory    | VideoUnmask, VideoUnmaskSwap, ButtonUnmask, ButtonUnmaskSwap         |
+| **Reference**  | Object memory     | PickHighlight, VideoRepick, VideoPlaceButton, VideoPlaceOrder         |
+| **Imitation**  | Procedural memory | MoveCube, InsertPeg, PatternLock, RouteStick                          |
+
+The dataset includes 1,600 demonstrations in total (100 per task).
+
+Our adaptation of MemoryVLA on RoboMME can be found [here](https://github.com/RoboMME/MemoryVLA).
+
+## Citation
+
+```bibtex
+@article{dai2026robomme,
+  title={RoboMME: Benchmarking and Understanding Memory for Robotic Generalist Policies},
+  author={Dai, Yinpei and Fu, Hongze and Lee, Jayjun and Liu, Yuejiang and Zhang, Haoran and Yang, Jianing and Finn, Chelsea and Fazeli, Nima and Chai, Joyce},
+  journal={arXiv preprint arXiv:2603.04639},
+  year={2026}
+}
+```