try

app.py

@@ -53,7 +53,7 @@ def process_image(image, text_input=None):
     image = Image.fromarray(image)  # Convert NumPy array to PIL Image
     task_prompt = '<DocVQA>'
     results = run_example(task_prompt, image, text_input)[task_prompt].replace("<pad>", "")
-    return results, None
+    return results
 
 
 css = """
@@ -74,7 +74,6 @@ with gr.Blocks(css=css) as demo:
                 submit_btn = gr.Button(value="Submit")
             with gr.Column():
                 output_text = gr.Textbox(label="Output Text")
-                output_img = gr.Image(label="Output Image")
 
         gr.Examples(
             examples=[
@@ -84,12 +83,12 @@ with gr.Blocks(css=css) as demo:
                 ["image.jpg", "What's the share of Industry Switchers Gained?"]
             ],
             inputs=[input_img, text_input],
-            outputs=[output_text, output_img],
+            outputs=[output_text],
             fn=process_image,
             cache_examples=True,
             label='Try examples'
         )
 
-        submit_btn.click(process_image, [input_img, text_input], [output_text, output_img])
+        submit_btn.click(process_image, [input_img, text_input], [output_text])
 
 demo.launch(debug=True)