Update audio_model.py

audio_model.py

@@ -14,6 +14,20 @@ speaker_embeddings = torch.tensor(embeddings_dataset[7440]["xvector"]).unsqueeze
 
 device = "cuda" if torch.cuda.is_available() else "cpu"
 
+replacements = [
+    ("á", "a"),
+    ("í", "i"),
+    ("ñ", "n"),
+    ("ó", "o"),
+    ("ú", "u"),
+    ("ü", "u"),
+]
+
+def cleanup_text(text):
+    for src, dst in replacements:
+        text = text.replace(src, dst)
+    return text
+    
 ### TEXT TO AUDIO SPEECH  MODEL 2    
 # Define la función que convierte texto en voz
 def synthesize_speech(text):