Improve model card: Add pipeline tag, library name, links, and description

README.md

@@ -1,7 +1,21 @@
 ---
-license: apache-2.0
-language:
-- en
 base_model:
 - Qwen/Qwen2.5-VL-7B-Instruct
----
+language:
+- en
+license: apache-2.0
+pipeline_tag: image-text-to-text
+library_name: transformers
+---
+
+# GeoVista: Web-Augmented Agentic Visual Reasoning for Geolocalization
+
+This repository hosts **GeoVista**, an agentic model designed for web-augmented visual reasoning, primarily focused on the geolocalization task. GeoVista leverages deep multimodal understanding by integrating tool invocation, such as an image-zoom-in tool and a web-search tool, within its reasoning loop. This allows the model to perform nuanced visual grounding and to confirm or refine hypotheses using external web information.
+
+GeoVista was introduced in the paper:
+[**GeoVista: Web-Augmented Agentic Visual Reasoning for Geolocalization**](https://huggingface.co/papers/2511.15705)
+
+## Links
+
+- **Project Page:** [https://ekonwang.github.io/geo-vista/](https://ekonwang.github.io/geo-vista/)
+- **Code Repository:** [https://github.com/ekonwang/GeoVista](https://github.com/ekonwang/GeoVista)