Update README.md

README.md

@@ -6,73 +6,85 @@ task_categories:
 language:
 - zh
 - en
-
-dataset_info:
-  - config_name: Full-page OCR
-    features:
-      - name: image_filename
-        dtype: string
-      - name: image
-        dtype: image
-      - name: label
-        dtype: string
-    splits:
-      - name: test
-
 configs:
-- config_name: Full-page OCR  # 例如：default
+- config_name: easy
+  data_files:
+  - split: test
+    path: data/easy/*.parquet
+- config_name: medium
   data_files:
   - split: test
-    path: 
-      - data/easy/*.parquet  # 指定easy子文件夹下所有.parquet文件
-      - data/medium/*.parquet  # 指定medium子文件夹下所有.parquet文件
-      - data/hard/*.parquet  # 指定hard子文件夹下所有.parquet文件
+    path: data/medium/*.parquet
+- config_name: hard
+  data_files:
+  - split: test
+    path: data/hard/*.parquet
 tags:
 - art
+- calligraphy
+- chinese
+- visual-language-model
+- document-understanding
 size_categories:
 - 1K<n<10K
-
 ---
 
-
 # 🧠 CalliReader: Contextualizing Chinese Calligraphy via an Embedding-aligned Vision Language Model
+
 <div align="center">
-  <a href="https://github.com/LoYuXr/CalliReader">📂 Code</a>
+  <a href="https://github.com/LoYuXr/CalliReader">📂 Code</a> •
   <a href="https://arxiv.org/pdf/2503.06472">📄 Paper</a>
 </div>
 
-**CalliBench** is aimed to comprehensively evaluate VLMs' performance on the recognition and understanding of Chinese calligraphy. 
+## Dataset Description
+
+**CalliBench** is a comprehensive benchmark designed to evaluate Vision-Language Models (VLMs) on the recognition and understanding of Chinese calligraphy.
+
+### Dataset Subsets
+
+CalliBench contains three distinct subsets organized by difficulty level:
 
+#### 🟢 Easy Subset
+- **Samples**: ~1,000 image-annotation pairs
+- **Characteristics**: Standard script styles, clear layouts, well-known works
+- **Use Case**: Basic calligraphy recognition and transcription
 
+#### 🟡 Medium Subset  
+- **Samples**: ~1,000 image-annotation pairs
+- **Characteristics**: Mixed script styles, moderate complexity, historical variants
+- **Use Case**: Intermediate-level analysis and contextual understanding
 
-## 📦 Dataset Summary
+#### 🔴 Hard Subset
+- **Samples**: ~1,192 image-annotation pairs
+- **Characteristics**: Cursive scripts, complex layouts, rare characters, damaged works
+- **Use Case**: Advanced contextual reasoning and expert-level interpretation
 
-* **Samples**: 3,192 image–annotation pairs
-* **Tasks**: **Full-page recognition** and **Contextual VQA** (choice of author/layout/style, bilingual interpretation, and intent analysis).
-* **Annotations**:
-  
-  * Metadata of author, layout, and style.
-  * Fine-grained annotations of **character-wise bounding boxes and labels**.
-  * Certain samples include **contextual VQA**.
+### Dataset Summary
 
-## 🧪 How To Use
-All **.parqeut** files of different tiers can be found in the sub-folders of **data**. **Pandas** can be used to parse and further process those files.
+- **Total Samples**: 3,192 image–annotation pairs
+- **Tasks**: 
+  - **Full-page recognition**: Complete transcription of calligraphy works
+  - **Contextual VQA**: Author identification, layout analysis, style classification, bilingual interpretation, and intent analysis
 
-For example, to load a sample and convert its image into a .jpg file: 
-```
+### Usage Examples
+
+#### Loading Specific Subsets
+
+```python
 import pandas as pd
 import io
 from PIL import Image
 
-df = pd.read_parquet('./data/hard/hard.parquet')
-
-image_data = df.iloc[0]['image'] 
-image = Image.open(io.BytesIO(image_data['bytes']))
-
-image.save('output_image.jpg') 
+# Load easy subset
+df_easy = pd.read_parquet('./data/easy/easy.parquet')
 
-```
+# Load medium subset  
+df_medium = pd.read_parquet('./data/medium/medium.parquet')
 
-## 🤗 License
+# Load hard subset
+df_hard = pd.read_parquet('./data/hard/hard.parquet')
 
-Apache 2.0 – open for research and commercial use.
+# Access sample data
+sample = df_hard.iloc[0]
+image_data = sample['image']
+image = Image.open(io.BytesIO(image_data['bytes']))