Update app.py

app.py

@@ -11,7 +11,7 @@ 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():
-    return manual_cap.update(value=None,placeholder=ph_message)
+    return manual_cap.update(value=None,placeholder=ph_message), caption_output.update(value=None)
 
 def infer(image_input, manual_caption, duration_in, seed):
     print(duration_in)
@@ -103,7 +103,7 @@ with gr.Blocks(css="style.css") as demo:
 
         gr.HTML(article)
 
-    input_img.change(clean, None, manual_cap)
+    input_img.change(clean, None, [manual_cap,caption_output])
     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)