# flake8: noqa: E402
import sys, os
import logging
logging.getLogger("numba").setLevel(logging.WARNING)
logging.getLogger("markdown_it").setLevel(logging.WARNING)
logging.getLogger("urllib3").setLevel(logging.WARNING)
logging.getLogger("matplotlib").setLevel(logging.WARNING)
logging.basicConfig(
level=logging.INFO, format="| %(name)s | %(levelname)s | %(message)s"
)
logger = logging.getLogger(__name__)
import torch
import argparse
import commons
import utils
from models import SynthesizerTrn
from text.symbols import symbols
from text import cleaned_text_to_sequence, get_bert
from text.cleaner import clean_text
import gradio as gr
import webbrowser
import numpy as np
net_g = None
if sys.platform == "darwin" and torch.backends.mps.is_available():
device = "mps"
os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1"
else:
device = "cpu"
def get_text(text, language_str, hps):
norm_text, phone, tone, word2ph = clean_text(text, language_str)
phone, tone, language = cleaned_text_to_sequence(phone, tone, language_str)
if hps.data.add_blank:
phone = commons.intersperse(phone, 0)
tone = commons.intersperse(tone, 0)
language = commons.intersperse(language, 0)
for i in range(len(word2ph)):
word2ph[i] = word2ph[i] * 2
word2ph[0] += 1
bert = get_bert(norm_text, word2ph, language_str, device)
del word2ph
assert bert.shape[-1] == len(phone), phone
if language_str == "ZH":
bert = bert
ja_bert = torch.zeros(768, len(phone))
elif language_str == "JP":
ja_bert = bert
bert = torch.zeros(1024, len(phone))
else:
bert = torch.zeros(1024, len(phone))
ja_bert = torch.zeros(768, len(phone))
assert bert.shape[-1] == len(
phone
), f"Bert seq len {bert.shape[-1]} != {len(phone)}"
phone = torch.LongTensor(phone)
tone = torch.LongTensor(tone)
language = torch.LongTensor(language)
return bert, ja_bert, phone, tone, language
def infer(text, sdp_ratio, noise_scale, noise_scale_w, length_scale, sid, language):
global net_g
bert, ja_bert, phones, tones, lang_ids = get_text(text, language, hps)
with torch.no_grad():
x_tst = phones.to(device).unsqueeze(0)
tones = tones.to(device).unsqueeze(0)
lang_ids = lang_ids.to(device).unsqueeze(0)
bert = bert.to(device).unsqueeze(0)
ja_bert = ja_bert.to(device).unsqueeze(0)
x_tst_lengths = torch.LongTensor([phones.size(0)]).to(device)
del phones
speakers = torch.LongTensor([hps.data.spk2id[sid]]).to(device)
audio = (
net_g.infer(
x_tst,
x_tst_lengths,
speakers,
tones,
lang_ids,
bert,
ja_bert,
sdp_ratio=sdp_ratio,
noise_scale=noise_scale,
noise_scale_w=noise_scale_w,
length_scale=length_scale,
)[0][0, 0]
.data.cpu()
.float()
.numpy()
)
del x_tst, tones, lang_ids, bert, x_tst_lengths, speakers
torch.cuda.empty_cache()
return audio
def tts_fn(
text, speaker, sdp_ratio, noise_scale, noise_scale_w, length_scale, language
):
print(f"text: {text}, speaker: {speaker}")
slices = text.split("\n")
audio_list = []
with torch.no_grad():
for slice in slices:
audio = infer(
slice,
sdp_ratio=sdp_ratio,
noise_scale=noise_scale,
noise_scale_w=noise_scale_w,
length_scale=length_scale,
sid=speaker,
language=language,
)
audio_list.append(audio)
silence = np.zeros(hps.data.sampling_rate) # 生成1秒的静音
audio_list.append(silence) # 将静音添加到列表中
audio_concat = np.concatenate(audio_list)
return "Success", (hps.data.sampling_rate, audio_concat)
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument(
"-m", "--model", default="./models/G_270K.pth", help="path of your model"
)
parser.add_argument(
"-c",
"--config",
default="./models/config.json",
help="path of your config file",
)
parser.add_argument(
"--share", default=False, help="make link public", action="store_true"
)
parser.add_argument(
"-d", "--debug", action="store_true", help="enable DEBUG-LEVEL log"
)
parser.add_argument(
"--info_md", default='./info.md', help="info markdown file"
)
args = parser.parse_args()
if args.debug:
logger.info("Enable DEBUG-LEVEL log")
logging.basicConfig(level=logging.DEBUG)
hps = utils.get_hparams_from_file(args.config)
device = (
"cuda:0"
if torch.cuda.is_available()
else (
"mps"
if sys.platform == "darwin" and torch.backends.mps.is_available()
else "cpu"
)
)
net_g = SynthesizerTrn(
len(symbols),
hps.data.filter_length // 2 + 1,
hps.train.segment_size // hps.data.hop_length,
n_speakers=hps.data.n_speakers,
**hps.model,
).to(device)
_ = net_g.eval()
_ = utils.load_checkpoint(args.model, net_g, None, skip_optimizer=True)
speaker_ids = hps.data.spk2id
speakers = list(speaker_ids.keys())
languages = ["JP"]
with gr.Blocks(title="Umamusume-DeBERTa-VITS2") as app:
with gr.Row():
with gr.Column():
text = gr.TextArea(
label="Text",
placeholder="Input Text Here",
value="サイーゲームス！。ウマ娘、プリティーダービー！",
)
speaker = gr.Dropdown(
choices=speakers, value=speakers[0], label="Speaker"
)
sdp_ratio = gr.Slider(
minimum=0, maximum=1, value=0.2, step=0.05, label="SDP Ratio (个人认为更大的SDP Ratio会产生\"感情更强烈的\"语音) (I personally believe that a larger SDP Ratio will make the generated voice \"emotionally stronger\")"
)
noise_scale = gr.Slider(
minimum=0.1, maximum=2, value=0.6, step=0.05, label="Noise Scale"
)
noise_scale_w = gr.Slider(
minimum=0.1, maximum=2, value=0.8, step=0.05, label="Noise Scale W"
)
length_scale = gr.Slider(
minimum=0.1, maximum=2, value=1, step=0.05, label="Length Scale (控制生成语音的长度) (Controlling the length of the generated audio)"
)
language = gr.Dropdown(
choices=languages, value=languages[0], label="Language"
)
btn = gr.Button("Generate!", variant="primary")
with gr.Column():
text_output = gr.Textbox(label="Message")
audio_output = gr.Audio(label="Output Audio")
samples = gr.Textbox(label="WEIRD Samples Given By GPT-4")
samples.value = "⚠ 强烈不建议将所有内容扔进输入，这会导致相当久的推理时间 ⚠\n" \
"⚠ すべての内容をお入りになることがお勧めしませんで、生成時間が非常に長くなるでしょう ⚠\n" \
"⚠ Throwing Everything into text input leads to unexpected long inference time ⚠\n" \
"おはよう、今日も一緒に頑張りましょうね！\n" \
+ "ねえねえ、あなたの好きなお料理作ってあげるよ！\n" \
+ "きゃー！びっくりさせないでよ～！\n" \
+ "あのね、新しいドレス買ったの。どう思う？\n" \
+ "あっ、遅くなっちゃった！ごめんなさい！\n" \
+ "今日のデート、すごく楽しかったよ！また行きましょうね！\n" \
+ "私のこと、好き？\n" \
+ "あなたといると、時間があっという間に過ぎちゃうね。\n" \
+ "あたし、あなたが大好きだよ。\n" \
+ "ねえ、もっと話して！あなたの話、大好きなの！\n" \
+ "あっ、それ可愛いね！私に似合うかな？\n" \
+ "あなたのこと、ずっと考えてたんだよ。\n" \
+ "今日はどんな一日だった？私にも話して！\n" \
+ "あなたの笑顔、大好き！もっと見せて！\n" \
+ "おやすみ、いい夢見てね！"
with open(args.info_md, 'r', encoding='UTF-8') as file:
data = file.read()
md_info = gr.Markdown(data)
btn.click(
tts_fn,
inputs=[
text,
speaker,
sdp_ratio,
noise_scale,
noise_scale_w,
length_scale,
language,
],
outputs=[text_output, audio_output],
)
app.launch()