Update app.py

app.py

@@ -127,8 +127,7 @@ def infer(genre_txt_content, lyrics_txt_content, num_segments, max_new_tokens):
         "--stage2_batch_size", "4",
         "--output_dir", f"{output_dir}",
         "--cuda_idx", "0",
-        "--max_new_tokens", str(max_new_tokens),
-        "--disable_offload_model"
+        "--max_new_tokens", str(max_new_tokens)
     ]
 
     # Set up environment variables for CUDA with optimized settings
@@ -222,13 +221,14 @@ with gr.Blocks() as demo:
                 
             with gr.Column():
                 
-                num_segments = gr.Number(label="Number of Segments", value=2, interactive=True)
+                num_segments = gr.Number(label="Number of Segments", value=2, interactive=False)
                 max_new_tokens = gr.Slider(label="Max New Tokens", minimum=500, maximum="3000", step=500, value=1500, interactive=True)
                 
                 submit_btn = gr.Button("Submit")
                 music_out = gr.Audio(label="Audio Result")
-                instrumental = gr.Audio(label="Intrumental")
-                vocal = gr.Audio(label="Vocal")
+                with gr.Accordion("Vocal & Instrumental", open=False):
+                    instrumental = gr.Audio(label="Intrumental")
+                    vocal = gr.Audio(label="Vocal")
 
         gr.Examples(
             examples = [