make_clone.py

#pip install "pydantic<2.0"
#pip install "numpy<=1.24.4"

import os
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
os.environ["TF_FORCE_CPU"] = "true"
import soundfile as sf
from TTS.api import TTS

tts = TTS(model_name="tts_models/multilingual/multi-dataset/xtts_v2")
reference_wav = f"ref.wav"
wav = tts.tts(
        text = "नमस्ते, आप कैसे हैं? मैं टैक्स ऑफिस से बोल रहा हूँ",
        speaker_wav= reference_wav,
        language= "hi"
    )
sf.write(f"clone_hindi.wav", wav, 22050)