Add model card and metadata

README.md

@@ -1,3 +1,30 @@
----
-license: mit
----
+---
+license: mit
+pipeline_tag: text-to-image
+---
+
+# FiMR: Enhanced Text-to-Image Generation by Fine-grained Multimodal Reasoning
+
+**FiMR** (Fine-grained Multimodal Reasoning) is a framework that enhances text-to-image generation in unified Multimodal Large Language Models (MLLMs). It leverages decomposed visual question answering (VQA) to break down an input prompt into minimal semantic units—such as entities and attributes—and verifies each unit to generate explicit, fine-grained feedback for targeted image refinement.
+
+[**Paper**](https://arxiv.org/abs/2604.13491) | [**Code**](https://github.com/KU-AGI/FiMR)
+
+## Overview
+Existing multimodal reasoning-based image generation methods often rely on holistic image-text alignment judgments. FiMR improves upon this by using decomposed VQA to verify detailed prompt attributes. This fine-grained self-reasoning and self-refinement enable MLLMs to achieve more precise improvements in image-prompt alignment and overall generation quality at test time.
+
+## Citation
+
+```bibtex
+@misc{kim2026enhancedtexttoimagegenerationfinegrained,
+      title={Enhanced Text-to-Image Generation by Fine-grained Multimodal Reasoning}, 
+      author={Yongjin Kim and Yoonjin Oh and Yerin Kim and Hyomin Kim and Jeeyoung Yun and Yujung Heo and Minjun Kim and Sungwoong Kim},
+      year={2026},
+      eprint={2604.13491},
+      archivePrefix={arXiv},
+      primaryClass={cs.CV},
+      url={https://arxiv.org/abs/2604.13491}, 
+}
+```
+
+## Acknowledgement
+The authors thank the developers of [Janus-Series](https://github.com/deepseek-ai/janus) and [Janus-Pro-R1](https://github.com/wendell0218/Janus-Pro-R1) for making their models and code publicly available.