Update app.py

app.py

@@ -146,7 +146,7 @@ def audio_tensor_to_opus_bytes(audio_tensor: torch.Tensor, sample_rate: int = 24
 
 
 # Initialize G2P for English (American)
-g2p = espeak.EspeakG2P(language="en-us")
+g2p = en.G2P(trf=False, british=False, fallback=None)
 
 def tokenizer(text: str):
     """
@@ -154,7 +154,7 @@ def tokenizer(text: str):
     """
     phonemes_string, _ = g2p(text)
     phonemes = [ph for ph in phonemes_string]
-    print(phonemes)
+    print(phonemes_string)
     tokens = [phoneme_vocab[phoneme] for phoneme in phonemes if phoneme in phoneme_vocab]
     return tokens