adjust layout.

app.py

@@ -136,10 +136,6 @@ examples = [
     ["house.png", "How could someone get out of the house?"],
     ["flower.jpg", "Question: What is this flower and where is it's origin? Answer:"],
     ["forbidden_city.webp", "In what dynasties was this place build?"],
-    # [
-    #     "sunset.png",
-    #     "Write a romantic message that goes along this photo.",
-    # ],
 ]
 
 with gr.Blocks() as iface:
@@ -152,97 +148,96 @@ with gr.Blocks() as iface:
         with gr.Column():
             image_input = gr.Image(type="pil")
 
-            with gr.Row():
-                sampling = gr.Radio(
-                    choices=["Beam search", "Nucleus sampling"],
-                    value="Beam search",
-                    label="Text Decoding Method",
-                    interactive=True,
-                )
-
-                temperature = gr.Slider(
-                    minimum=0.5,
-                    maximum=1.0,
-                    value=0.8,
-                    step=0.1,
-                    interactive=True,
-                    label="Temperature (used with nucleus sampling)",
-                )
-
-                len_penalty = gr.Slider(
-                    minimum=-1.0,
-                    maximum=2.0,
-                    value=1.0,
-                    step=0.2,
-                    interactive=True,
-                    label="Length Penalty (set to larger for longer sequence, used with beam search)",
-                )
-
-                rep_penalty = gr.Slider(
-                    minimum=1.0,
-                    maximum=5.0,
-                    value=1.5,
-                    step=0.5,
-                    interactive=True,
-                    label="Repeat Penalty (larger value prevents repetition)",
-                )
-
-            with gr.Row():
-                caption_output = gr.Textbox(lines=2, label="Caption Output (from OPT)")
-                caption_button = gr.Button(
-                    value="Caption it!", interactive=True, variant="primary"
-                )
-                caption_button.click(
-                    inference_caption,
-                    [
-                        image_input,
-                        sampling,
-                        temperature,
-                        len_penalty,
-                        rep_penalty,
-                    ],
-                    [caption_output],
-                )
+            # with gr.Row():
+            sampling = gr.Radio(
+                choices=["Beam search", "Nucleus sampling"],
+                value="Beam search",
+                label="Text Decoding Method",
+                interactive=True,
+            )
+
+            temperature = gr.Slider(
+                minimum=0.5,
+                maximum=1.0,
+                value=1.0,
+                step=0.1,
+                interactive=True,
+                label="Temperature (used with nucleus sampling)",
+            )
+
+            len_penalty = gr.Slider(
+                minimum=-1.0,
+                maximum=2.0,
+                value=1.0,
+                step=0.2,
+                interactive=True,
+                label="Length Penalty (set to larger for longer sequence, used with beam search)",
+            )
+
+            rep_penalty = gr.Slider(
+                minimum=1.0,
+                maximum=5.0,
+                value=1.5,
+                step=0.5,
+                interactive=True,
+                label="Repeat Penalty (larger value prevents repetition)",
+            )
 
         with gr.Column():
-            chat_input = gr.Textbox(lines=2, label="Chat Input")
-
-            with gr.Row():
-                chatbot = gr.Chatbot(label="Chat Output (from FlanT5)")
-                image_input.change(lambda: (None, "", "", []), [], [chatbot, chat_input, caption_output, state])
-
-            with gr.Row():
-
-                clear_button = gr.Button(value="Clear", interactive=True)
-                clear_button.click(
-                    lambda: ("", None, [], []),
-                    [],
-                    [chat_input, image_input, chatbot, state],
-                )
-
-                submit_button = gr.Button(
-                    value="Submit", interactive=True, variant="primary"
-                )
-                submit_button.click(
-                    inference_chat,
-                    [
-                        image_input,
-                        chat_input,
-                        sampling,
-                        temperature,
-                        len_penalty,
-                        rep_penalty,
-                        state,
-                    ],
-                    [chatbot, state],
-                )
+
+            with gr.Column():
+                # with gr.Row():
+                    caption_output = gr.Textbox(lines=1, label="Caption Output (from OPT)")
+                    caption_button = gr.Button(
+                        value="Caption it!", interactive=True, variant="primary"
+                    )
+                    caption_button.click(
+                        inference_caption,
+                        [
+                            image_input,
+                            sampling,
+                            temperature,
+                            len_penalty,
+                            rep_penalty,
+                        ],
+                        [caption_output],
+                    )
+
+            with gr.Column():
+                chat_input = gr.Textbox(lines=1, label="Chat Input")
+                with gr.Row():
+                    chatbot = gr.Chatbot(label="Chat Output (from FlanT5)")
+                    image_input.change(lambda: (None, "", "", []), [], [chatbot, chat_input, caption_output, state])
+
+                with gr.Row():
+
+                    clear_button = gr.Button(value="Clear", interactive=True)
+                    clear_button.click(
+                        lambda: ("", None, [], []),
+                        [],
+                        [chat_input, image_input, chatbot, state],
+                    )
+
+                    submit_button = gr.Button(
+                        value="Submit", interactive=True, variant="primary"
+                    )
+                    submit_button.click(
+                        inference_chat,
+                        [
+                            image_input,
+                            chat_input,
+                            sampling,
+                            temperature,
+                            len_penalty,
+                            rep_penalty,
+                            state,
+                        ],
+                        [chatbot, state],
+                    )
 
     examples = gr.Examples(
         examples=examples,
         inputs=[image_input, chat_input],
-#         outputs=[chatbot, state],
-#         run_on_click=True,
-#         fn = inference_chat,
     )
 
 iface.queue(concurrency_count=1, api_open=False, max_size=10)