update layout

app.py

@@ -349,6 +349,7 @@ with gr.Blocks() as demo:
             <br/>
         """)
 
+    banner_1 = gr.Markdown(value="# Text-to-image Generation")
     with gr.Row():
         with gr.Column():
             text_prompt_t2i = gr.Textbox(
@@ -387,15 +388,20 @@ with gr.Blocks() as demo:
                          [text_prompt_t2i, guidance_scale_t2i, generation_timesteps_t2i],
                          [generated_img_t2i])
 
+    banner_2 = gr.Markdown(value="# Text-guided inpainting")
     with gr.Row():
         inpainting_input_img = gr.Image(
             label="Input image",
             type="pil",
+            height=256,
+            width=256,
         )
         inpainting_input_mask = gr.Image(
             label="Inpainting mask",
             image_mode="L",
             type="pil",
+            height=256,
+            width=256,
         )
 
         with gr.Column():
@@ -443,6 +449,7 @@ with gr.Blocks() as demo:
                                  guidance_scale_inpainting, generation_timesteps_inpainting],
                                 [generated_img_inpainting])
 
+    banner_3 = gr.Markdown(value="# Text-guided extrapolation")
     with gr.Row():
         extra_input_img = gr.Image(
             label="Input image",
@@ -510,6 +517,8 @@ with gr.Blocks() as demo:
                                 [extra_input_img, text_prompt_extrapolation, left_extrapolation, right_extrapolation,
                                  guidance_scale_extrapolation, generation_timesteps_extrapolation],
                                 [generated_img_extrapolation])
+
+    banner_4 = gr.Markdown(value="# Multimodal understanding")
     with gr.Row():
         with gr.Row():
             chat_input_img = gr.Image(