Add task categories, language and Arxiv ID to metadata

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

This PR improves the dataset card by adding:
- `task_categories`: `zero-shot-image-classification` and `image-to-text` to improve discoverability.
- `language`: `zh` (Chinese), as this is a Chinese-centric vision-language dataset.
- `arxiv_id`: `2601.10305` to link the paper directly to the dataset repository.
- Links to the project page and GitHub repository at the top of the README.
- A quick sample usage section for the `datasets` library.

These changes help researchers find, identify, and use the dataset more efficiently.

README.md

@@ -1,25 +1,32 @@
 ---
+license: cc-by-4.0
+task_categories:
+- zero-shot-image-classification
+- image-to-text
+language:
+- zh
+arxiv_id: 2601.10305
 dataset_info:
   features:
-    - name: url
-      dtype: string
-    - name: alt_text
-      dtype: string
-    - name: recaption
-      dtype: string
+  - name: url
+    dtype: string
+  - name: alt_text
+    dtype: string
+  - name: recaption
+    dtype: string
   splits:
-    - name: train
-      num_bytes: 19795236
-      num_examples: 99892381
+  - name: train
+    num_bytes: 19795236
+    num_examples: 99892381
   download_size: 19795236
   dataset_size: 99892381
 configs:
-  - config_name: default
-    data_files:
-      - split: train
-        path: data/train-*
-license: cc-by-4.0
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
 ---
+
 <div align="center">
 
 <img src="Figures/danqing.svg" width="30%">
@@ -28,6 +35,8 @@ license: cc-by-4.0
 
 <h1 align="center">DanQing: An Up-to-Date Large-Scale Chinese Vision-Language Pre-training Dataset</h1>
 
+[**Project Page**](https://deepglint.github.io/DanQing/) | [**Paper**](https://huggingface.co/papers/2601.10305) | [**Code**](https://github.com/deepglint/DanQing)
+
 </div>
 
 
@@ -60,6 +69,18 @@ Hengyu Shen<sup>∗</sup>, [**Tiancheng Gu**](https://scholar.google.com/citatio
 
 ---
 
+## 🚀 Sample Usage
+
+You can load the metadata (URLs and captions) directly using the `datasets` library:
+
+```python
+from datasets import load_dataset
+
+ds = load_dataset("DeepGlint-AI/DanQing100M")
+```
+
+---
+
 ## 📑 Table of Contents
 - [💡 Highlights](#-highlights)
 - [💻 Dataset Information](#-dataset-information)
@@ -286,5 +307,4 @@ If you find this repository useful, please use the following BibTeX entry for ci
 
 ### ⭐ Don't forget to star this repository if you find it helpful!
 
-</div>
-
+</div>