update ui.

app.py

@@ -125,13 +125,6 @@ def inference_caption(
     return output[0]
 
 
-# def clear_fn(image_input, chatbot, chat_input, caption_output, state):
-#     # if image_input is None:
-#     return (None, "", "", [])
-# else:
-#     return chatbot, chat_input, caption_output, state
-
-
 title = """<h1 align="center">BLIP-2</h1>"""
 description = """Gradio demo for BLIP-2, image-to-text generation from Salesforce Research. To use it, simply upload your image, or click one of the examples to load them.
 <br> <strong>Disclaimer</strong>: This is a research prototype and is not intended for production use. No data including but not restricted to text and images is collected."""
@@ -149,7 +142,9 @@ examples = [
     ["forbidden_city.webp", "In what dynasties was this place built?"],
 ]
 
-with gr.Blocks() as iface:
+with gr.Blocks(
+    css=""".message.svelte-w6rprc.svelte-w6rprc.svelte-w6rprc {font-size: 20px; margin-top: 20px}"""
+) as iface:
     state = gr.State([])
 
     gr.Markdown(title)
@@ -216,8 +211,12 @@ with gr.Blocks() as iface:
 
             gr.Markdown("""Trying prompting your input for chat; e.g. example prompt for QA, \"Question: {} Answer:\"""")
             with gr.Row():
-                with gr.Column(scale=1.5):
-                    chatbot = gr.Chatbot(label="Chat Output (from FlanT5)")
+                with gr.Column(
+                    scale=1.5, 
+                ):
+                    chatbot = gr.Chatbot(
+                        label="Chat Output (from FlanT5)",
+                    )
 
                 # with gr.Row():
                 with gr.Column(scale=1):