app.py

import os
import gradio as gr
from TTS.api import TTS
import torch
import numpy as np
os.environ['COQUI_TOS_AGREED'] = '1'
device = "cuda" if torch.cuda.is_available() else "cpu"
# 初始化 TTS 模型
tts = TTS("tts_models/multilingual/multi-dataset/xtts_v2").to(device)

# 定义语音合成函数
def text_to_speech(text):
    # 将文本转换为语音
    #wav = tts.tts(text)
    # 保存为临时文件
    wav_file = "output.wav"
    wav = tts.tts(text, language = "zh-cn", speaker_wav = "E_sample.wav")
    return wav, tts.sample_rate

# 创建 Gradio 接口
iface = gr.Interface(
    fn=text_to_speech,
    inputs=gr.Textbox(lines=2, placeholder="输入文本..."),
    outputs=gr.Audio(type="numpy"),
    title=f"测试",
    description=f"当前device: {device}"
)

# 启动接口
iface.launch()