Add model card and metadata for DSR Suite model

Hi! I'm Niels from the Hugging Face community science team. I've opened this PR to add a comprehensive model card and metadata for your DSR Suite model.

This PR:
- Adds the `video-text-to-text` pipeline tag for better discoverability.
- Adds the `library_name: transformers` tag as indicated by the `config.json` and `tokenizer_config.json` files, enabling automated code snippets.
- Links the model to the paper [Learning to Reason in 4D: Dynamic Spatial Understanding for Vision Language Models](https://huggingface.co/papers/2512.20557).
- Includes links to the project's GitHub repository, the associated Hugging Face dataset, and collection.
- Provides an introduction summarizing the model's capabilities in dynamic spatial reasoning and details on its usage for evaluation.
- Includes the correct BibTeX citation and acknowledgements.

Please review and merge if this looks good to you!

README.md

@@ -1,4 +1,63 @@
----
-license: apache-2.0
-extra_gated_eu_disallowed: true
----
+---
+license: apache-2.0
+extra_gated_eu_disallowed: true
+pipeline_tag: video-text-to-text
+library_name: transformers
+base_model: Qwen/Qwen2.5-VL-7B-Instruct
+tags:
+- spatial-reasoning
+- 4d-vision
+- vlm
+---
+
+# Learning to Reason in 4D: Dynamic Spatial Understanding for Vision Language Models
+
+This repository contains the model weights for the **DSR Suite**, which introduces advancements in dynamic spatial reasoning for Vision Language Models (VLMs), as presented in the paper [Learning to Reason in 4D: Dynamic Spatial Understanding for Vision Language Models](https://huggingface.co/papers/2512.20557).
+
+## Introduction
+Vision-language models (VLMs) typically excel at general understanding but demonstrate weaknesses in **Dynamic Spatial Reasoning (DSR)** – the ability to reason about the evolution of object geometry and relationships in 3D space over time. To address this gap, we introduce **DSR Suite**, which comprises:
+
+1.  **Automated Data Generation Pipeline**: A system that constructs multiple-choice question-answer pairs from in-the-wild videos for DSR.
+2.  **DSR-Train**: A training dataset of 50K QAs generated by the pipeline.
+3.  **DSR-Bench**: A human-refined benchmark with 1484 QAs for rigorous evaluation.
+4.  **Geometry Selection Module (GSM)**: A lightweight module designed to seamlessly integrate geometric priors from 3D foundation models into VLMs, specifically a **Qwen2.5-VL-7B** backbone, without compromising general understanding capabilities.
+
+Experiments show that integrating DSR-Train and GSM into Qwen2.5-VL-7B significantly enhances its dynamic spatial reasoning.
+
+## Resources
+- **Paper**: [Learning to Reason in 4D: Dynamic Spatial Understanding for Vision Language Models](https://huggingface.co/papers/2512.20557)
+- **GitHub Repository**: [https://github.com/TencentARC/DSR_Suite](https://github.com/TencentARC/DSR_Suite)
+- **Hugging Face Dataset**: [TencentARC/DSR_Suite-Data](https://huggingface.co/datasets/TencentARC/DSR_Suite-Data)
+- **Hugging Face Collection**: [TencentARC/dsr-suite](https://huggingface.co/collections/TencentARC/dsr-suite)
+
+## Usage and Evaluation
+For detailed instructions on environment setup, data generation, model training, and benchmark evaluation, please refer to the official [DSR_Suite GitHub repository](https://github.com/TencentARC/DSR_Suite).
+
+The evaluation framework is based on [VLMEvalKit](https://github.com/open-compass/VLMEvalKit). An example command for evaluating a trained model (like `Qwen2.5-VL-7B-Instruct-ForVideo-Spatial`) on the `Spatial-Reasoning` task is:
+
+```bash
+cd VLMEvalKit_mine
+CUDA_VISIBLE_DEVICES=0 python run.py --data Spatial-Reasoning --model Qwen2.5-VL-7B-Instruct-ForVideo-Spatial --work-dir spatial_reasoning
+```
+
+## Citation
+If you find our work useful, please consider citing:
+
+```bibtex
+@misc{zhou2025learning,
+      title={Learning to Reason in 4D: Dynamic Spatial Understanding for Vision Language Models}, 
+      author={Shengchao Zhou, Yuxin Chen, Yuying Ge, Wei Huang, Jiehong Lin, Ying Shan, Xiaojuan Qi},
+      year={2025},
+      eprint={2512.20557},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/2512.20557}, 
+}
+```
+
+## Acknowledgement
+This work builds upon the following projects:
+- [Qwen2.5-VL](https://github.com/QwenLM/Qwen3-VL): The model codebase we built upon.
+- [VLMEvalKit](https://github.com/open-compass/VLMEvalKit): The evaluation framework we built upon.
+- [Grounded SAM2](https://github.com/IDEA-Research/Grounded-SAM-2), [Orient Anything](https://github.com/SpatialVision/Orient-Anything), [π^3](https://github.com/yyfz/Pi3): Models used in our data generation pipeline to extract 3D cues.
+- [Koala-36M](https://github.com/KlingTeam/Koala-36M): The video database we build QAs upon.