Update app.py

app.py

@@ -187,12 +187,13 @@ def inference_click(image_input, point_prompt, click_mode, enable_wiki, language
 
     click_state[0] = input_points
     click_state[1] = input_labels
+    state = state + [("Image point: {}, Input label: {}".format(click_state[0], click_state[1]), None)]
 
     return state, click_state
 
 def submit_caption(image_input, enable_wiki, language, sentiment, factuality, length, image_embedding, state, click_state, original_size, input_size, text_refiner, visual_chatgpt):
-    coordinate = json.dumps(list(zip(click_state[0], click_state[1])))
-    prompt = get_click_prompt(coordinate, click_state, 'Single')
+    coordinates = json.dumps(list(zip(click_state[0], click_state[1])))
+    prompt = get_click_prompt(coordinates, click_state, 'Single')
     input_points = prompt['input_point']
     input_labels = prompt['input_label']
 
@@ -216,15 +217,16 @@ def submit_caption(image_input, enable_wiki, language, sentiment, factuality, le
     enable_wiki = True if enable_wiki in ['True', 'TRUE', 'true', True, 'Yes', 'YES', 'yes'] else False
     out = model.inference(image_input, prompt, controls, disable_gpt=True, enable_wiki=enable_wiki, verbose=True, args={'clip_filter': False})[0]
 
-    state = state + [("Image point: {}, Input label: {}".format(prompt["input_point"], prompt["input_label"]), None)]
     state = state + [(None, "raw_caption: {}".format(out['generated_captions']['raw_caption']))]
-    update_click_state(click_state, out['generated_captions']['raw_caption'], click_mode)
+
+    update_click_state(click_state, out['generated_captions']['raw_caption'], 'Single')
     text = out['generated_captions']['raw_caption']
     input_mask = np.array(out['mask'].convert('P'))
     image_input = mask_painter(np.array(image_input), input_mask)
     origin_image_input = image_input
     image_input = create_bubble_frame(image_input, text, (input_points[-1][0], input_points[-1][1]), input_mask,
                                       input_points=input_points, input_labels=input_labels)
+
     if visual_chatgpt is not None:
         new_image_path = get_new_image_name('chat_image', func_name='upload')
         image_input.save(new_image_path)
@@ -246,9 +248,10 @@ def submit_caption(image_input, enable_wiki, language, sentiment, factuality, le
                                                   input_mask,
                                                   input_points=input_points, input_labels=input_labels)
         return state, state, click_state, refined_image_input
-    
+
     return state, state, click_state, image_input
 
+
 def get_sketch_prompt(mask: Image.Image):
     """
     Get the prompt for the sketcher.
@@ -556,7 +559,7 @@ def create_ui():
         submit_button_click.click(submit_caption,
                                   [origin_image, enable_wiki, language, sentiment, factuality, length, image_embedding, state, click_state, original_size, input_size, text_refiner, visual_chatgpt],
                                   [chatbot, state, click_state, image_input])
-
+        
         image_input.clear(
             lambda: (None, [], [], [[], [], []], "", "", ""),
             [],