Update model card metadata and pipeline tag

This PR improves the model card's metadata by:
- Updating the `pipeline_tag` to `image-text-to-text` for better discoverability.
- Adding `library_name: transformers` as the model architecture is compatible with the Transformers library.
- Adding the `arxiv` metadata tag to link the model with its research paper: [MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methods](https://huggingface.co/papers/2601.21821).

README.md

@@ -1,11 +1,14 @@
 ---
-license: apache-2.0
 datasets:
 - OpenDataArena/MMFineReason-1.8M
 language:
 - en
-pipeline_tag: visual-question-answering
+license: apache-2.0
+pipeline_tag: image-text-to-text
+library_name: transformers
+arxiv: 2601.21821
 ---
+
 <div align="center">
   <h1>MMFineReason</h1>
   <p><strong>Closing the Multimodal Reasoning Gap via Open Data-Centric Methods</strong></p>
@@ -25,7 +28,7 @@ pipeline_tag: visual-question-answering
 </figure>
 
 ---
-This repository provides **MMFineReason-8B**; detailed dataset information is available at https://huggingface.co/datasets/OpenDataArena/MMFineReason-1.8M.
+This repository provides **MMFineReason-8B**, introduced in the paper [MMFineReason: Closing the Multimodal Reasoning Gap via Open Data-Centric Methods](https://huggingface.co/papers/2601.21821). Detailed dataset information is available at https://huggingface.co/datasets/OpenDataArena/MMFineReason-1.8M.
 
 ## 📖 Overview
 
@@ -95,7 +98,7 @@ SFT drives major gains in mathematical reasoning (e.g., MathVision: 53.9% → 67
 
 ## 📄 License
 
-This dataset is released under the [Apache 2.0 License](https://opensource.org/licenses/Apache-2.0). Individual source datasets may have their own licenses.
+This dataset and model are released under the [Apache 2.0 License](https://opensource.org/licenses/Apache-2.0). Individual source datasets may have their own licenses.
 
 ---