Update app.py

app.py

@@ -621,9 +621,8 @@ def infer_example(image_path, audio_path, text, num_steps, raw_image_path, sessi
     result = infer(image_path, audio_path, text, [[720, 400]], num_steps, session_id, progress)
 
     return result
-    
-@spaces.GPU(duration=get_duration)
-def infer(image_path, audio_path, text, orientation_state, num_steps, session_id = None, progress=gr.Progress(track_tqdm=True),):
+
+def infer_scene(image_path, audio_path, text, orientation_state, num_steps, session_id = None, progress=gr.Progress(track_tqdm=True),):
 
     if image_path is None:
 
@@ -636,7 +635,17 @@ def infer(image_path, audio_path, text, orientation_state, num_steps, session_id
     if session_id is None:
         session_id = uuid.uuid4().hex
 
-        
+    try:
+        result = infer(image_path, audio_path, text, orientation_state, num_steps, session_id, progress)
+    except Exception as e:
+        err = str(e).lower()
+        print(f"{session_id} failed due to {err}")        
+
+    return result
+    
+@spaces.GPU(duration=get_duration)
+def infer(image_path, audio_path, text, orientation_state, num_steps, session_id = None, progress=gr.Progress(track_tqdm=True),):
+
     output_dir = os.path.join(os.environ["PROCESSED_RESULTS"], session_id)
 
     audio_dir = output_dir + '/audio'
@@ -946,7 +955,7 @@ with gr.Blocks(css=css) as demo:
                     )
 
     infer_btn.click(
-        fn=infer,
+        fn=infer_scene,
         inputs=[image_input, audio_input, text_input, orientation_state, num_steps, session_state],
         outputs=[output_video]
     )