change whisper version

app.py

@@ -78,7 +78,7 @@ def process_audio(input_audio, new_chunk):
     stream, transcription = transcribe(input_audio, new_chunk)
     text = autocomplete(transcription)
     print (transcription, text)
-    api = TTS(model_name="tts_models/fra/fairseq/vits").to(device)
+    api = TTS(model_name="tts_models/multilingual/multi-dataset/xtts_v2").to(device)
     api.tts_to_file(text, file_path="output.wav", split_sentences=True)
     audio = "./output.wav"
     return stream, text, audio

requirements.txt

@@ -5,5 +5,4 @@ Cython
 spacy==3.7
 torchaudio
 transformers
-tts
-espeakng
+tts