modify app

app.py

@@ -41,6 +41,9 @@ def process_audio(input_audio, reference_audio):
     elif output_audio.ndim > 2:
         output_audio = output_audio.squeeze()
     
+    print(output_audio.shape)
+    print(param_output)
+    
     return (sr, output_audio), param_output
 
 def perform_ito(input_audio, reference_audio, ito_reference_audio, num_steps, optimizer, learning_rate, af_weights):
@@ -96,7 +99,7 @@ with gr.Blocks() as demo:
         process_button = gr.Button("Process Mastering Style Transfer")
         
         with gr.Row():
-            output_audio = gr.Audio(label="Output Audio")
+            output_audio = gr.Audio(label="Output Audio", type='tuple')
             param_output = gr.Textbox(label="Predicted Parameters", lines=10)
 
         process_button.click(

inference.py

@@ -349,8 +349,3 @@ if __name__ == "__main__":
         args.input_path, args.reference_path, args.ito_reference_path, ito_config, args.perform_ito
     )
 
-    # Save the output audio
-    sf.write("output_mastered.wav", output_audio.T, sr)
-    if ito_output_audio is not None:
-        sf.write("ito_output_mastered.wav", ito_output_audio.T, sr)
-