Update app.py

app.py

@@ -1,14 +1,23 @@
+import spaces
 import torch
 import gradio as gr
 from chatterbox_vc import VoiceConverter
 
-device = "cuda" if torch.cuda.is_available() else "cpu"
+vc = None
 
-print(f"Using device: {device}")
 
-vc = VoiceConverter(device=device)
+def get_converter():
+    global vc
 
+    if vc is None:
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+        print(f"Loading ChatterboxVC on {device}...")
+        vc = VoiceConverter(device=device)
 
+    return vc
+
+
+@spaces.GPU
 def convert(source_audio, target_audio):
     if source_audio is None:
         raise gr.Error("Please upload a source audio file.")
@@ -18,7 +27,9 @@ def convert(source_audio, target_audio):
 
     output = "output.wav"
 
-    vc.convert(
+    converter = get_converter()
+
+    converter.convert(
         source_audio,
         target_audio,
         output,
@@ -54,4 +65,5 @@ with gr.Blocks() as demo:
         outputs=output,
     )
 
+
 demo.launch()