update space

app.py

@@ -10,7 +10,8 @@ os.environ['REGIONAL_POOL'] = '2x'
 os.environ['FORCE_NO_DOWNSAMPLE'] = '1'
 os.environ['LOAD_VISION_EARLY'] = '1'
 os.environ['SKIP_LOAD_VIT'] = '1'
-    
+
+import spaces
 
 import gradio as gr
 import torch
@@ -146,6 +147,7 @@ def extract_audio(videos_file_path):
     my_clip = mp.VideoFileClip(videos_file_path)
     return my_clip.audio
 
+@spaces.GPU(duration=120)
 def ola_inference(multimodal, audio_path):
     visual, text = multimodal["files"][0], multimodal["text"]
     if visual.endswith("image2.png"):
@@ -353,4 +355,4 @@ demo = gr.Interface(
     article=bibtext,
 )
 # Launch the Gradio app
-demo.launch(server_name="0.0.0.0",server_port=80)
+demo.launch()