Update app.py

app.py

@@ -1,10 +1,10 @@
 import streamlit as st
 import torch
 import numpy as np
-from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
 import pyaudio
-import sounddevice as sd
-from TTS.api import TTS
+from transformers import Wav2Vec2Processor, Wav2Vec2ForCTC
+from gtts import gTTS
+import os
 
 class VoiceAssistant:
     def __init__(self):
@@ -12,9 +12,6 @@ class VoiceAssistant:
         self.processor = Wav2Vec2Processor.from_pretrained("facebook/wav2vec2-large-xlsr-53-spanish")
         self.model = Wav2Vec2ForCTC.from_pretrained("facebook/wav2vec2-large-xlsr-53-spanish")
         
-        # Cargar modelo TTS para síntesis de voz en español (modelo corregido)
-        self.tts_model = TTS(model_name="microsoft/speecht5_tts", progress_bar=False)
-        
         # Parámetros de audio
         self.sample_rate = 16000
         self.chunk_size = 480
@@ -66,8 +63,10 @@ class VoiceAssistant:
         return "Respuesta generada para: " + text
     
     def text_to_speech(self, text):
-        output_path = "response.wav"
-        self.tts_model.tts_to_file(text=text, file_path=output_path)
+        # Usar gTTS para convertir texto a voz
+        tts = gTTS(text=text, lang='es')
+        output_path = "response.mp3"
+        tts.save(output_path)
         return output_path
     
     def run(self):