|
import gradio as gr |
|
import os |
|
os.system('cd monotonic_align && python setup.py build_ext --inplace && cd ..') |
|
|
|
import time |
|
import json |
|
import math |
|
import torch |
|
from torch import nn |
|
from torch.nn import functional as F |
|
from torch.utils.data import DataLoader |
|
import re |
|
import langid |
|
import jieba |
|
import commons |
|
import utils |
|
from data_utils import TextAudioLoader, TextAudioCollate, TextAudioSpeakerLoader, TextAudioSpeakerCollate |
|
from models import SynthesizerTrn |
|
from text.symbols import symbols |
|
from text import text_to_sequence, cleaned_text_to_sequence |
|
from text.cleaners import japanese_cleaners |
|
from scipy.io.wavfile import write |
|
|
|
def getMixText(text): |
|
langid.set_languages(['zh','en']) |
|
seg_list = jieba.cut(text, cut_all=False) |
|
clean_list=[] |
|
for seg in seg_list: |
|
langtext='[ZH]' |
|
if(len(seg)>0): |
|
lang=langid.classify(seg)[0] |
|
if lang == 'en': |
|
langtext='[EN]' |
|
elif lang=='zh': |
|
langtext='[ZH]' |
|
clean_list.append(langtext+seg+langtext) |
|
return ''.join(clean_list) |
|
|
|
def get_text(text, hps): |
|
text_norm = text_to_sequence(text, hps.data.text_cleaners) |
|
if hps.data.add_blank: |
|
text_norm = commons.intersperse(text_norm, 0) |
|
text_norm = torch.LongTensor(text_norm) |
|
|
|
return text_norm |
|
|
|
hps_ms = utils.get_hparams_from_file("save_model/config.json") |
|
hps = utils.get_hparams_from_file("save_model/config.json") |
|
net_g_ms = SynthesizerTrn( |
|
len(symbols), |
|
hps_ms.data.filter_length // 2 + 1, |
|
hps_ms.train.segment_size // hps.data.hop_length, |
|
n_speakers=hps_ms.data.n_speakers, |
|
**hps_ms.model) |
|
|
|
npclists=[] |
|
with open("save_model/npclists.txt",'r') as r: |
|
for npc in r.readlines(): |
|
npclists.append(npc.split('|')[-1]) |
|
print(npc) |
|
r.close |
|
|
|
def tts(spkid, text): |
|
if(len(re.findall(r'\[ZH\].*?\[ZH\]', text))==0 and len(re.findall(r'\[EN\].*?\[EN\]', text))==0): |
|
text=getMixText(text) |
|
sid = torch.LongTensor([spkid]) |
|
stn_tst = get_text(text, hps_ms) |
|
|
|
with torch.no_grad(): |
|
x_tst = stn_tst.unsqueeze(0) |
|
x_tst_lengths = torch.LongTensor([stn_tst.size(0)]) |
|
t1 = time.time() |
|
audio = net_g_ms.infer(x_tst, x_tst_lengths, sid=sid, noise_scale=.667, noise_scale_w=0.8, length_scale=1)[0][ |
|
0, 0].data.float().numpy() |
|
t2 = time.time() |
|
return "成功,耗时"+str((t2-t1))+"s", (hps.data.sampling_rate, audio) |
|
|
|
|
|
_ = utils.load_checkpoint("save_model/model.pth", net_g_ms, None) |
|
|
|
def clean_text(text): |
|
return japanese_cleaners(text) |
|
|
|
app = gr.Blocks() |
|
with app: |
|
with gr.Tabs(): |
|
with gr.TabItem("Basic"): |
|
tts_input1 = gr.TextArea(label="在这输入文字", value="基于VITS的中英混合语音合成模型，当前进度为45epoch,30000 Steps，正在持续训练中。。") |
|
tts_input2 = gr.Dropdown(label="人物", choices=npclists, type="index", value=npclists[0]) |
|
tts_submit = gr.Button("合成", variant="primary") |
|
tts_output1 = gr.Textbox(label="信息") |
|
tts_output2 = gr.Audio(label="结果") |
|
tts_submit.click(tts, [tts_input2, tts_input1], [tts_output1, tts_output2]) |
|
app.launch() |
