Add robotics pipeline tag, library name, and paper metadata

Hi! I'm Niels from the Hugging Face community science team.

I've opened this PR to update the model card with standard metadata to improve its discoverability and integration on the Hub:
- Added `pipeline_tag: robotics` to categorize the model.
- Added `library_name: transformers` as the card identifies it as the supported runtime engine.
- Linked the model to the official research paper via the `arxiv` field.
- Specified `license: other` for the custom NVIDIA NSCLv1 license.
- Added a BibTeX citation for researchers using this work.

Please review and merge if this looks good to you!

README.md

@@ -1,7 +1,19 @@
 ---
 base_model:
 - nvidia/Cosmos-Predict2-2B-Video2World
+pipeline_tag: robotics
+library_name: transformers
+license: other
+tags:
+- robotics
+- nvidia
+- cosmos-policy
+- robot-manipulation
+- visuo-motor-control
+- imitation-learning
+arxiv: 2601.16163
 ---
+
 # **Cosmos-Policy-RoboCasa-Predict2-2B**
 
 [**Cosmos Policy**](https://huggingface.co/collections/nvidia/cosmos-policy) | [**Code**](http://github.com/NVlabs/cosmos-policy) | [**White Paper**](https://arxiv.org/abs/2601.16163)| [**Website**](https://research.nvidia.com/labs/dir/cosmos-policy/)
@@ -237,4 +249,11 @@ Please report security vulnerabilities or NVIDIA AI Concerns [here](https://www.
 
 If you use this model, please cite the Cosmos Policy paper:
 
-(Cosmos Policy BibTeX citation coming soon!)
+```bibtex
+@article{kim2026cosmos,
+  title={Cosmos Policy: Fine-Tuning Video Models for Visuomotor Control and Planning},
+  author={Kim, Moo Jin and Gao, Yihuai and Lin, Tsung-Yi and Lin, Yen-Chen and Ge, Yunhao and Lam, Grace and Liang, Percy and Song, Shuran and Liu, Ming-Yu and Finn, Chelsea and Gu, Jinwei},
+  journal={arXiv preprint arXiv:2601.16163},
+  year={2026}
+}
+```