Upload app.py

app.py

@@ -9,10 +9,10 @@ if __name__ == '__main__':
     dtype = "float16" if torch.cuda.is_available() else "float32"
     
     # You can change the model ID here based on what you want to deploy:
-    # "Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice"
-    # "Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign"
-    # "Qwen/Qwen3-TTS-12Hz-1.7B-Base"
-    model_id = "Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice"
+    # "Qwen/Qwen3-TTS-12Hz-1.7B-CustomVoice" (Standard voice only)
+    # "Qwen/Qwen3-TTS-12Hz-1.7B-VoiceDesign" (Design your own voice)
+    # "Qwen/Qwen3-TTS-12Hz-1.7B-Base" (Includes 0-shot Voice Cloning)
+    model_id = "Qwen/Qwen3-TTS-12Hz-1.7B-Base"
     
     # Construct the arguments expected by the qwen-tts-demo CLI
     sys.argv = [
@@ -23,9 +23,5 @@ if __name__ == '__main__':
         "--ip", "0.0.0.0", 
         "--port", "7860" # Standard Hugging Face Spaces Gradio port
     ]
-    
-    # Nvidia T4 GPUs (Turing architecture) do not support Flash Attention 2
-    # Flash Attention 2 requires Ampere GPUs or newer.
-    sys.argv.append("--no-flash-attn")
         
     sys.exit(main())