Update app.py

app.py

@@ -16,6 +16,7 @@ from fastrtc import (
 )
 from huggingface_hub import InferenceClient
 from transformers import AutoModelForSpeechSeq2Seq, AutoProcessor, pipeline
+import spaces
 
 device = "cuda:0" if torch.cuda.is_available() else "cpu"
 torch_dtype = torch.float16 if torch.cuda.is_available() else torch.float32
@@ -49,7 +50,7 @@ tts_model = get_tts_model()
 
 conversations: dict[str, list[dict[str, str]]] = {}
 
-
+@spaces.GPU
 def response(
     audio: tuple[int, np.ndarray],
     hf_token: str | None,