Update app.py

app.py

@@ -7,16 +7,15 @@ import torch
 # Load JinaAI CLIP model
 model = AutoModel.from_pretrained("jinaai/jina-clip-v1", trust_remote_code=True)
 
-# Function to process input (convert to text or PIL image)
+# Function to process input (convert text or image)
 def process_input(input_data, input_type):
     if input_type == "Text":
         return model.encode_text([input_data]) if input_data.strip() else None
     elif input_type == "Image":
-        if isinstance(input_data, np.ndarray):  # If it's a NumPy array (Gradio default)
-            image = Image.fromarray(input_data)
-        else:
-            return None  # Invalid input type
-        return model.encode_image([image])
+        if isinstance(input_data, np.ndarray):  # Gradio provides image as NumPy array
+            image = Image.fromarray(input_data)  # Convert NumPy array to PIL Image
+            return model.encode_image([image])
+        return None  # Invalid input type
     return None
 
 # Function to compute similarity