Update app.py

app.py

@@ -71,20 +71,23 @@ def encode_image_to_base64(image_array):
     return base64.b64encode(buffered.getvalue()).decode('utf-8')
 
 
-def format_query_for_model(text_input, model_type="owlv2"):
+# def format_query_for_model(text_input, model_type="owlv2"):
     """Format query based on model requirements"""
-    # Extract objects (e.g., "count frogs and horses" -> ["frog", "horse"])
+    # Extract objects (e.g., "detect a lion" -> "lion")
     text = text_input.lower()
     words = [w.strip('.,?!') for w in text.split() 
              if w not in ['count', 'detect', 'show', 'me', 'the', 'and', 'a', 'an']]
     
     if model_type == "owlv2":
-        print([["a photo of " + obj for obj in words]])
-        return [["a photo of " + obj for obj in words]]
+        # Return just the list of queries for Owlv2, not nested list
+        queries = ["a photo of " + obj for obj in words]
+        print("Owlv2 queries:", queries)
+        return queries
     else:  # DINO
-        # DINO only works with single object queries with format "a object."
-        print(f"a {words[0]}.")
-        return f"a {words[0]}."
+        # DINO query format
+        query = f"a {words[0]}."
+        print("DINO query:", query)
+        return query
 
 def detect_objects(query_text):
     if state.current_image is None:
@@ -94,6 +97,7 @@ def detect_objects(query_text):
     draw = ImageDraw.Draw(image)
     
     if state.current_model == "owlv2":
+        # For Owlv2, pass the text queries directly
         inputs = owlv2_processor(text=query_text, images=image, return_tensors="pt").to(device)
         with torch.no_grad():
             outputs = owlv2_model(**inputs)
@@ -101,6 +105,7 @@ def detect_objects(query_text):
             outputs=outputs, threshold=0.2, target_sizes=torch.Tensor([image.size[::-1]])
         )
     else:  # DINO
+        # For DINO, pass the single text query
         inputs = dino_processor(images=image, text=query_text, return_tensors="pt").to(device)
         with torch.no_grad():
             outputs = dino_model(**inputs)
@@ -125,7 +130,6 @@ def detect_objects(query_text):
         "message": f"Detected {len(boxes)} objects"
     }
 
-
 def identify_plant():
     if state.current_image is None:
         return {"error": "No image provided"}