add interface logic

app.py

@@ -1,7 +1,6 @@
 import gradio as gr
 import torch
 from PIL import Image, ImageDraw, ImageFont 
-import requests
 from transformers import AutoProcessor
 from modeling_florence2 import Florence2ForConditionalGeneration
 from configuration_florence2 import Florence2Config
@@ -79,6 +78,7 @@ def process_image(image, task):
     
     return parsed_answer
 
+
 def main_process(image, task):
     result = process_image(image, task)
     
@@ -88,12 +88,12 @@ def main_process(image, task):
         else:
             fig = plot_bbox(image, result[TASK_PROMPTS[task]])
             output_image = fig_to_pil(fig)
-        return output_image, None
+        return output_image, gr.update(visible=True), None, gr.update(visible=False)
     else:
-        return None, str(result)
+        return None, gr.update(visible=False), str(result), gr.update(visible=True)
 
 def reset_outputs():
-    return None, None
+    return None, gr.update(visible=False), None, gr.update(visible=False)
 
 with gr.Blocks(title="Florence-2 Demo") as iface:
     gr.Markdown("# Florence-2 Demo")
@@ -110,23 +110,16 @@ with gr.Blocks(title="Florence-2 Demo") as iface:
     output_image = gr.Image(label="Processed Image", visible=False)
     output_text = gr.Textbox(label="Output", visible=False)
     
-    def process_and_show(image, task):
-        img_output, text_output = main_process(image, task)
-        if img_output is not None:
-            return {output_image: gr.update(value=img_output, visible=True), output_text: gr.update(visible=False)}
-        else:
-            return {output_image: gr.update(visible=False), output_text: gr.update(value=text_output, visible=True)}
-    
     submit_button.click(
-        fn=process_and_show,
+        fn=main_process,
         inputs=[image_input, task_dropdown],
-        outputs=[output_image, output_text]
+        outputs=[output_image, output_image, output_text, output_text]
     )
     
     reset_button.click(
         fn=reset_outputs,
         inputs=[],
-        outputs=[output_image, output_text]
+        outputs=[output_image, output_image, output_text, output_text]
     )
 
 iface.launch()