compatible with cpu instance

app.py

@@ -114,7 +114,12 @@ def text2speech(target_text, game, speaker):
         # Convert  token <|s_23456|> to int 23456 
         speech_tokens = extract_speech_ids(speech_tokens)
 
-        speech_tokens = torch.tensor(speech_tokens).cuda().unsqueeze(0).unsqueeze(0)
+        if torch.cuda.is_available():
+            # Move speech tokens to GPU
+            speech_tokens = torch.tensor(speech_tokens).cuda().unsqueeze(0).unsqueeze(0)
+        else:
+            # If CUDA is not available, keep on CPU
+            speech_tokens = torch.tensor(speech_tokens).unsqueeze(0).unsqueeze(0)
 
         # Decode the speech tokens to speech waveform
         gen_wav = Codec_model.decode_code(speech_tokens)