Update app.py

app.py

@@ -15,9 +15,10 @@ from mel_processing import spectrogram_torch
 
 limitation = os.getenv("SYSTEM") == "spaces"  # limit text and audio length in huggingface spaces
 
+device = 'cpu'
 
-def get_text(text, hps, is_phoneme):
-    text_norm = text_to_sequence(text, hps.symbols, [] if is_phoneme else hps.data.text_cleaners)
+def get_text(text, hps):
+    text_norm = text_to_sequence(text, hps.data.text_cleaners)
     if hps.data.add_blank:
         text_norm = commons.intersperse(text_norm, 0)
     text_norm = LongTensor(text_norm)
@@ -51,9 +52,6 @@ def create_tts_fn(model, hps, speaker_ids):
     return tts_fn
 
 
-
-
-
 def create_to_phoneme_fn(hps):
     def to_phoneme_fn(text):
         return _clean_text(text, hps.data.text_cleaners) if text != "" else ""
@@ -95,7 +93,7 @@ if __name__ == '__main__':
         hps.data.filter_length // 2 + 1,
         hps.train.segment_size // hps.data.hop_length,
         n_speakers=hps.data.n_speakers,
-        **hps.model)
+        **hps.model.to(device))
     utils.load_checkpoint(model_path, model, None)
     model.eval()
     speaker_ids = [sid for sid, name in enumerate(hps.speakers) if name != "None"]