Update app.py

app.py

@@ -38,51 +38,7 @@ if example_dir is not None:
             video_examples.append([example_file])
 
 
-with gr.Blocks() as interface:
-    gr.Markdown(HEADER)
-    with gr.Row():
-        chatbot = gr.Chatbot(type="messages", elem_id="chatbot", height=710)
-
-        with gr.Column():
-            with gr.Tab(label="Input"):
-
-                with gr.Row():
-                    input_video = gr.Video(sources=["upload"], label="Upload Video")
-                    input_image = gr.Image(sources=["upload"], type="filepath", label="Upload Image")
-
-                if len(image_examples):
-                    gr.Examples(image_examples, inputs=[input_image], label="Example Images")
-                if len(video_examples):
-                    gr.Examples(video_examples, inputs=[input_video], label="Example Videos")
-
-                input_text = gr.Textbox(label="Input Text", placeholder="Type your message here and press enter to submit")
-
-                submit_button = gr.Button("Generate")
-
-            with gr.Tab(label="Configure"):
-                with gr.Accordion("Generation Config", open=True):
-                    do_sample = gr.Checkbox(value=True, label="Do Sample")
-                    temperature = gr.Slider(minimum=0.0, maximum=1.0, value=0.2, label="Temperature")
-                    top_p = gr.Slider(minimum=0.0, maximum=1.0, value=0.9, label="Top P")
-                    max_new_tokens = gr.Slider(minimum=0, maximum=4096, value=2048, step=1, label="Max New Tokens")
-
-                with gr.Accordion("Video Config", open=True):
-                    fps = gr.Slider(minimum=0.0, maximum=10.0, value=1, label="FPS")
-                    max_frames = gr.Slider(minimum=0, maximum=256, value=180, step=1, label="Max Frames")
-
-    input_video.change(_on_video_upload, [chatbot, input_video], [chatbot, input_video])
-    input_image.change(_on_image_upload, [chatbot, input_image], [chatbot, input_image])
-    input_text.submit(_on_text_submit, [chatbot, input_text], [chatbot, input_text])
-    submit_button.click(
-        _predict,
-        [
-            chatbot, input_text, do_sample, temperature, top_p, max_new_tokens,
-            fps, max_frames
-        ],
-        [chatbot],
-    )
-
-    def _on_video_upload(messages, video):
+def _on_video_upload(messages, video):
         if video is not None:
             # messages.append({"role": "user", "content": gr.Video(video)})
             messages.append({"role": "user", "content": {"path": video}})
@@ -162,5 +118,50 @@ with gr.Blocks() as interface:
             yield messages
 
 
+with gr.Blocks() as interface:
+    gr.Markdown(HEADER)
+    with gr.Row():
+        chatbot = gr.Chatbot(type="messages", elem_id="chatbot", height=710)
+
+        with gr.Column():
+            with gr.Tab(label="Input"):
+
+                with gr.Row():
+                    input_video = gr.Video(sources=["upload"], label="Upload Video")
+                    input_image = gr.Image(sources=["upload"], type="filepath", label="Upload Image")
+
+                if len(image_examples):
+                    gr.Examples(image_examples, inputs=[input_image], label="Example Images")
+                if len(video_examples):
+                    gr.Examples(video_examples, inputs=[input_video], label="Example Videos")
+
+                input_text = gr.Textbox(label="Input Text", placeholder="Type your message here and press enter to submit")
+
+                submit_button = gr.Button("Generate")
+
+            with gr.Tab(label="Configure"):
+                with gr.Accordion("Generation Config", open=True):
+                    do_sample = gr.Checkbox(value=True, label="Do Sample")
+                    temperature = gr.Slider(minimum=0.0, maximum=1.0, value=0.2, label="Temperature")
+                    top_p = gr.Slider(minimum=0.0, maximum=1.0, value=0.9, label="Top P")
+                    max_new_tokens = gr.Slider(minimum=0, maximum=4096, value=2048, step=1, label="Max New Tokens")
+
+                with gr.Accordion("Video Config", open=True):
+                    fps = gr.Slider(minimum=0.0, maximum=10.0, value=1, label="FPS")
+                    max_frames = gr.Slider(minimum=0, maximum=256, value=180, step=1, label="Max Frames")
+
+    input_video.change(_on_video_upload, [chatbot, input_video], [chatbot, input_video])
+    input_image.change(_on_image_upload, [chatbot, input_image], [chatbot, input_image])
+    input_text.submit(_on_text_submit, [chatbot, input_text], [chatbot, input_text])
+    submit_button.click(
+        _predict,
+        [
+            chatbot, input_text, do_sample, temperature, top_p, max_new_tokens,
+            fps, max_frames
+        ],
+        [chatbot],
+    )
+
+
 if __name__ == "__main__":
     interface.launch()