updated readme

README.md

@@ -1,3 +1,32 @@
 ---
+language: en
+tags:
+- vision
+- multimodal
+- id-card-recognition
 license: apache-2.0
 ---
+
+# Qwen2-VL Persian ID Card Reader
+
+Fine-tuned version of Qwen2-VL for extracting information from Persian national ID cards.
+
+## Usage
+
+```python
+from transformers import Qwen2VLForConditionalGeneration, AutoProcessor
+from util.vision_util import process_vision_info
+
+model = Qwen2VLForConditionalGeneration.from_pretrained("your-username/qwen2-vl-idcard-reader", device_map="auto")
+processor = AutoProcessor.from_pretrained("your-username/qwen2-vl-idcard-reader")
+
+messages = [{
+    "role": "user",
+    "content": [
+        {"type": "image", "image": "id_card.jpg"},
+        {"type": "text", "text": "Extract ID card info as JSON..."}
+    ]
+}]
+
+# Process and run inference (as in your original code)
+...