Update app.py

app.py

@@ -11,15 +11,19 @@ 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 input_changes(input_img):
+    sound_output.update(visible=False)
+    
     if input_img == None:
         return manual_cap.update(value="",placeholder=ph_message), caption_output.update(value=None), sound_output.update(value=None)
     else:
         cap = caption(input_img, fn_index=0)
         print("gpt2 caption: '" + cap + "' • ")
         ph_update = "gpt2 caption: '" + cap + "' • "
+        
         return manual_cap.update(value="",placeholder=f"{ph_update}{ph_message}"), caption_output.update(value=cap), sound_output.update(value=None)
  
 def infer(image_input, manual_caption, duration_in, seed, caption_output):
+    sound_output.update(visible=True)
     print(duration_in)
     if manual_caption == "":
         cap = caption_output
@@ -100,7 +104,7 @@ with gr.Blocks(css="style.css") as demo:
                 seed_in = gr.Slider(label="Seed", value=440, minimum=45, maximum=10000, step=1)
         
         caption_output = gr.Textbox(label="Caption", visible=False, elem_id="text-caption")
-        sound_output = gr.Audio(label="Result", elem_id="sound-output")
+        sound_output = gr.Audio(label="Result", visible=False, elem_id="sound-output")
         
         generate = gr.Button("Generate SFX from Image")