Update README.md

README.md

@@ -34,7 +34,6 @@ The JSON file includes three fields:
 This task requires the model to predict the **description** of a widget based on its given **bbox**. Different models support different bbox formats:  
 - **Qwen model:** Prefers `[xmin, ymin, xmax, ymax]` (raw pixel coordinates)  
 - **Intern model:** Prefers `[xmin, ymin, xmax, ymax]` (normalized coordinates)  
-- **AndesVL-4B-UI model:** Prefers `<|box_start|>(xmin,ymin),(xmax,ymax)<|box_end|>`  (based on normalized coordinates multiplied by 10,000)
 
 **Accuracy metric:** A prediction (`pred`) is considered correct if its longest common substring with the ground-truth `description` is non-empty.  
 
@@ -47,7 +46,6 @@ The JSON file includes three fields:
 This task is the inverse of Referring: the model must predict the **bbox** of a widget given its **description**. Different models output different formats:  
 - Some models directly predict `[x_center, y_center]` (center coordinates)  
 - Others predict `[xmin, ymin, xmax, ymax]` (full bbox)  
-- **AndesVL-4B-UI** predicts `<|box_start|>(xmin,ymin),(xmax,ymax)<|box_end|>`  
 
 If the model outputs a full bbox, we compute its geometric center as `[x_center, y_center]`.  
 **Accuracy metric:** The predicted center must lie inside the ground-truth bbox.