Update app.py

app.py

@@ -1,4 +1,3 @@
-
 import streamlit as st
 import base64
 import io
@@ -46,16 +45,7 @@ def format_prompt(message, history):
     prompt += f"[INST] {message} [/INST]"
     return prompt
 
-def text_to_speech(text, speed=1.3):
-    tts = gTTS(text=text, lang='es')
-    audio_fp = io.BytesIO()
-    tts.write_to_fp(audio_fp)
-    audio_fp.seek(0)
-    return audio_fp
-
 def generate(audio_text, history, temperature=None, max_new_tokens=512, top_p=0.95, repetition_penalty=1.0):
-    client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")
-
     temperature = float(temperature) if temperature is not None else 0.9
     if temperature < 1e-2:
         temperature = 1e-2
@@ -81,18 +71,24 @@ def generate(audio_text, history, temperature=None, max_new_tokens=512, top_p=0.
     audio_file = text_to_speech(response, speed=1.3)
     return response, audio_file
 
-def main():
-    st.title("Grabación de Audio y Reconocimiento de Voz")
+def text_to_speech(text, speed=1.3):
+    tts = gTTS(text=text, lang='es')
+    audio_fp = io.BytesIO()
+    tts.write_to_fp(audio_fp)
+    audio_fp.seek(0)
+    return audio_fp
 
+def main():
+    audio_data = audiorecorder("Habla para grabar", "Deteniendo la grabación...")
+    
     if "history" not in st.session_state:
         st.session_state.history = []
 
-    audio_data = audiorecorder("Habla para grabar", "Deteniendo la grabación...")
-
     if not audio_data.empty():
-        st.audio(audio_data.export().read(), format="audio/wav")
-        audio_data.export("audio.wav", format="wav")
-        st.write(f"Frame rate: {audio_data.frame_rate}, Frame width: {audio_data.frame_width}, Duration: {audio_data.duration_seconds} seconds")
+
+        with st.spinner("Procesando..."):
+            st.audio(audio_data.export().read(), format="audio/wav")
+            audio_data.export("audio.wav", format="wav")
 
         audio_text = recognize_speech("audio.wav")