Update app.py

app.py

@@ -11,9 +11,10 @@ audio_gen = gr.Blocks.load(name="spaces/fffiloni/audioldm-text-to-audio-generati
 ph_message="If you're not happy with sound result, you can manually describe the scene depicted in your image :)"
 
 def clean():
-    manual_cap.update(value="",placeholder=ph_message)
-    caption_output.update(value=None)
-    sound_output.update(value=None)
+    cleaned_outputs = [manual_cap.update(value="",placeholder=ph_message),
+        caption_output.update(value=None),
+        sound_output.update(value=None)]
+    return cleaned_outputs
 
 def infer(image_input, manual_caption, duration_in, seed):
     print(duration_in)
@@ -105,7 +106,8 @@ with gr.Blocks(css="style.css") as demo:
 
         gr.HTML(article)
 
-    input_img.change(clean)
+    clean_out = [manual_cap, caption_output, sound_output]
+    input_img.change(clean, None, clean_out)
     generate.click(infer, inputs=[input_img, manual_cap, duration_in, seed_in], outputs=[caption_output, sound_output, manual_cap, share_group], api_name="i2fx")
     share_button.click(None, [], [], _js=share_js)