import sys
sys.path.append('third_party/Matcha-TTS')
from vllm import ModelRegistry
from cosyvoice.vllm.cosyvoice2 import CosyVoice2ForCausalLM
ModelRegistry.register_model("CosyVoice2ForCausalLM", CosyVoice2ForCausalLM)

from cosyvoice.cli.cosyvoice import CosyVoice2
from cosyvoice.utils.file_utils import load_wav
from cosyvoice.utils.common import set_all_random_seed
from tqdm import tqdm


def main():
    cosyvoice = CosyVoice2('pretrained_models/CosyVoice2-0.5B', load_jit=True, load_trt=True, load_vllm=True, fp16=True)
    prompt_speech_16k = load_wav('./asset/zero_shot_prompt.wav', 16000)
    for i in tqdm(range(100)):
        set_all_random_seed(i)
        for _, _ in enumerate(cosyvoice.inference_zero_shot('收到好友从远方寄来的生日礼物，那份意外的惊喜与深深的祝福让我心中充满了甜蜜的快乐，笑容如花儿般绽放。', '希望你以后能够做的比我还好呦。', prompt_speech_16k, stream=False)):
            continue


if __name__ == '__main__':
    main()