blackteay/GPT-SoVITS
1
0
Fork 0
mirror of https://github.com/RVC-Boss/GPT-SoVITS.git synced 2025-10-07 15:19:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add files via upload

Browse Source
This commit is contained in:
YYuX-1145 2024-03-23 18:03:18 +08:00 committed by GitHub
parent ed75ecdd6d
commit f4d34c390e
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 648 additions and 134 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

363
GPT_SoVITS/api_simple.py Normal file
View File

@ -0,0 +1,363 @@
"""
# api.py usage
` python api.py -dr "123.wav" -dt "一二三。" -dl "zh" `
## 执行参数:
`-s` - `SoVITS模型路径, 可在 config.py 中指定`
`-g` - `GPT模型路径, 可在 config.py 中指定`
调用请求缺少参考音频时使用
`-dr` - `默认参考音频路径`
`-dt` - `默认参考音频文本`
`-dl` - `默认参考音频语种, "中文","英文","日文","zh","en","ja"`
`-d` - `推理设备, "cuda","cpu"`
`-a` - `绑定地址, 默认"127.0.0.1"`
`-p` - `绑定端口, 默认9880, 可在 config.py 中指定`
`-fp` - `覆盖 config.py 使用全精度`
`-hp` - `覆盖 config.py 使用半精度`
`-hb` - `cnhubert路径`
`-b` - `bert路径`
## 调用:
### 推理
endpoint: `/`
使用执行参数指定的参考音频:
GET:
`http://127.0.0.1:9880?text=先帝创业未半而中道崩殂，今天下三分，益州疲弊，此诚危急存亡之秋也。&text_language=zh`
POST:
```json
{
"text": "先帝创业未半而中道崩殂，今天下三分，益州疲弊，此诚危急存亡之秋也。",
"text_language": "zh"
}
```
手动指定当次推理所使用的参考音频:
GET:
`http://127.0.0.1:9880?refer_wav_path=123.wav&prompt_text=一二三。&prompt_language=zh&text=先帝创业未半而中道崩殂，今天下三分，益州疲弊，此诚危急存亡之秋也。&text_language=zh`
POST:
```json
{
"refer_wav_path": "123.wav",
"prompt_text": "一二三。",
"prompt_language": "zh",
"text": "先帝创业未半而中道崩殂，今天下三分，益州疲弊，此诚危急存亡之秋也。",
"text_language": "zh"
}
```
RESP:
成功: 直接返回 wav 音频流， http code 200
失败: 返回包含错误信息的 json, http code 400
### 更换默认参考音频
endpoint: `/change_refer`
key与推理端一样
GET:
`http://127.0.0.1:9880/change_refer?refer_wav_path=123.wav&prompt_text=一二三。&prompt_language=zh`
POST:
```json
{
"refer_wav_path": "123.wav",
"prompt_text": "一二三。",
"prompt_language": "zh"
}
```
RESP:
成功: json, http code 200
失败: json, 400
### 命令控制
endpoint: `/control`
command:
"restart": 重新运行
"exit": 结束运行
GET:
`http://127.0.0.1:9880/control?command=restart`
POST:
```json
{
"command": "restart"
}
```
RESP: 无
"""
import argparse
import os
import sys
now_dir = os.getcwd()
sys.path.append(now_dir)
sys.path.append("%s/GPT_SoVITS" % (now_dir))
import soundfile as sf
from fastapi import FastAPI, Request, HTTPException
from fastapi.responses import StreamingResponse, JSONResponse
import uvicorn
from io import BytesIO
import inference_webui
from inference_webui import inference as get_tts_wav
import signal
import config as global_config
g_config = global_config.Config()
# AVAILABLE_COMPUTE = "cuda" if torch.cuda.is_available() else "cpu"
parser = argparse.ArgumentParser(description="GPT-SoVITS api")
parser.add_argument("-s", "--sovits_path", type=str, default=g_config.sovits_path, help="SoVITS模型路径")
parser.add_argument("-g", "--gpt_path", type=str, default=g_config.gpt_path, help="GPT模型路径")
parser.add_argument("-dr", "--default_refer_path", type=str, default="", help="默认参考音频路径")
parser.add_argument("-dt", "--default_refer_text", type=str, default="", help="默认参考音频文本")
parser.add_argument("-dl", "--default_refer_language", type=str, default="", help="默认参考音频语种")
parser.add_argument("-d", "--device", type=str, default=g_config.infer_device, help="cuda / cpu")
parser.add_argument("-a", "--bind_addr", type=str, default="0.0.0.0", help="default: 0.0.0.0")
parser.add_argument("-p", "--port", type=int, default=g_config.api_port, help="default: 9880")
#parser.add_argument("-fp", "--full_precision", action="store_true", default=False, help="覆盖config.is_half为False, 使用全精度")
#parser.add_argument("-hp", "--half_precision", action="store_true", default=False, help="覆盖config.is_half为True, 使用半精度")
# bool值的用法为 `python ./api.py -fp ...`
# 此时 full_precision==True, half_precision==False
parser.add_argument("-hb", "--hubert_path", type=str, default=g_config.cnhubert_path, help="覆盖config.cnhubert_path")
parser.add_argument("-b", "--bert_path", type=str, default=g_config.bert_path, help="覆盖config.bert_path")
args = parser.parse_args()
sovits_path = args.sovits_path
gpt_path = args.gpt_path
def change_sovits_weights(sovits_path):
if sovits_path is not None and sovits_path !="":
inference_webui.tts_pipline.init_vits_weights(sovits_path)
def change_gpt_weights(gpt_path):
if gpt_path is not None and gpt_path !="":
inference_webui.tts_pipline.init_t2s_weights(gpt_path)
change_sovits_weights(sovits_path)
change_gpt_weights(gpt_path)
class DefaultRefer:
def __init__(self, path, text, language):
self.path = args.default_refer_path
self.text = args.default_refer_text
self.language = args.default_refer_language
def is_ready(self) -> bool:
return is_full(self.path, self.text, self.language)
default_refer = DefaultRefer(args.default_refer_path, args.default_refer_text, args.default_refer_language)
device = args.device
port = args.port
host = args.bind_addr
def is_empty(*items): # 任意一项不为空返回False
for item in items:
if item is not None and item != "":
return False
return True
def is_full(*items): # 任意一项为空返回False
for item in items:
if item is None or item == "":
return False
return True
dict_language = {
"中文": "zh",
"英文": "en",
"日文": "ja",
"ZH": "zh",
"EN": "en",
"JA": "ja",
"zh": "zh",
"en": "en",
"ja": "ja"
}
def handle_control(command):
if command == "restart":
os.execl(g_config.python_exec, g_config.python_exec, *sys.argv)
elif command == "exit":
os.kill(os.getpid(), signal.SIGTERM)
exit(0)
def handle_change(path, text, language):
if is_empty(path, text, language):
return JSONResponse({"code": 400, "message": '缺少任意一项以下参数: "path", "text", "language"'}, status_code=400)
if path != "" or path is not None:
default_refer.path = path
if text != "" or text is not None:
default_refer.text = text
if language != "" or language is not None:
default_refer.language = language
print(f"[INFO] 当前默认参考音频路径: {default_refer.path}")
print(f"[INFO] 当前默认参考音频文本: {default_refer.text}")
print(f"[INFO] 当前默认参考音频语种: {default_refer.language}")
print(f"[INFO] is_ready: {default_refer.is_ready()}")
return JSONResponse({"code": 0, "message": "Success"}, status_code=200)
def handle(text, text_language,
refer_wav_path, prompt_text,
prompt_language, top_k,
top_p, temperature,
text_split_method, batch_size,
speed_factor, ref_text_free,
split_bucket,fragment_interval,
seed):
if (
refer_wav_path == "" or refer_wav_path is None
or prompt_text == "" or prompt_text is None
or prompt_language == "" or prompt_language is None
):
refer_wav_path, prompt_text, prompt_language = (
default_refer.path,
default_refer.text,
default_refer.language,
)
if not default_refer.is_ready():
return JSONResponse({"code": 400, "message": "未指定参考音频且接口无预设"}, status_code=400)
prompt_text = prompt_text.strip("\n")
prompt_language, text = prompt_language, text.strip("\n")
gen = get_tts_wav(text, text_language,
refer_wav_path, prompt_text,
prompt_language, top_k,
top_p, temperature,
text_split_method, batch_size,
speed_factor, ref_text_free,
split_bucket,fragment_interval,
seed
)
audio,_ = next(gen)
sampling_rate,audio_data=audio
wav = BytesIO()
sf.write(wav, audio_data, sampling_rate, format="wav")
wav.seek(0)
return StreamingResponse(wav, media_type="audio/wav")
app = FastAPI()
#clark新增-----2024-02-21
#可在启动后动态修改模型，以此满足同一个api不同的朗读者请求
@app.post("/set_model")
async def set_model(request: Request):
json_post_raw = await request.json()
global gpt_path
gpt_path=json_post_raw.get("gpt_model_path")
global sovits_path
sovits_path=json_post_raw.get("sovits_model_path")
print("gptpath"+gpt_path+";vitspath"+sovits_path)
change_sovits_weights(sovits_path)
change_gpt_weights(gpt_path)
return "ok"
# 新增-----end------
@app.post("/control")
async def control(request: Request):
json_post_raw = await request.json()
return handle_control(json_post_raw.get("command"))
@app.get("/control")
async def control(command: str = None):
return handle_control(command)
@app.post("/change_refer")
async def change_refer(request: Request):
json_post_raw = await request.json()
return handle_change(
json_post_raw.get("refer_wav_path"),
json_post_raw.get("prompt_text"),
json_post_raw.get("prompt_language")
)
@app.get("/change_refer")
async def change_refer(
refer_wav_path: str = None,
prompt_text: str = None,
prompt_language: str = None
):
return handle_change(refer_wav_path, prompt_text, prompt_language)
'''
@app.post("/")
async def tts_endpoint(request: Request):
json_post_raw = await request.json()
return handle(
json_post_raw.get("refer_wav_path"),
json_post_raw.get("prompt_text"),
json_post_raw.get("prompt_language"),
json_post_raw.get("text"),
json_post_raw.get("text_language"),
)
'''
@app.get("/")
async def tts_endpoint(
refer_wav_path: str = None,
prompt_text: str = None,
prompt_language: str = None,
text: str = None,
text_language: str = None,
top_k:int =5,
top_p:float =1,
temperature:float=1,
text_split_method:str="凑四句一切",
batch_size:int=20,
speed_factor:float=1,
ref_text_free:bool=False,
split_bucket:bool=True,
fragment_interval:float=0.3,
seed:int=-1,
):
return handle(text, text_language,
refer_wav_path, prompt_text,
prompt_language, top_k,
top_p, temperature,
text_split_method, batch_size,
speed_factor, ref_text_free,
split_bucket,fragment_interval,
seed)
if __name__ == "__main__":
uvicorn.run(app, host=host, port=port, workers=1)

419
GPT_SoVITS/inference_webui.py
View File

@ -1,17 +1,20 @@
'''
"""
按中英混合识别
按日英混合识别
多语种启动切分识别语种
全部按中文识别
全部按英文识别
全部按日文识别
'''
"""
import random
import os, sys
now_dir = os.getcwd()
sys.path.append(now_dir)
import os, re, logging
logging.getLogger("markdown_it").setLevel(logging.ERROR)
logging.getLogger("urllib3").setLevel(logging.ERROR)
logging.getLogger("httpcore").setLevel(logging.ERROR)
@ -34,7 +37,7 @@ gpt_path = os.environ.get("gpt_path", None)
sovits_path = os.environ.get("sovits_path", None)
cnhubert_base_path = os.environ.get("cnhubert_base_path", None)
bert_path = os.environ.get("bert_path", None)
import gradio as gr
from TTS_infer_pack.TTS import TTS, TTS_Config
from TTS_infer_pack.text_segmentation_method import get_method
@ -50,18 +53,18 @@ if torch.cuda.is_available():
# device = "mps"
else:
device = "cpu"
dict_language = {
i18n("中文"): "all_zh",#全部按中文识别
i18n("英文"): "en",#全部按英文识别#######不变
i18n("日文"): "all_ja",#全部按日文识别
i18n("中英混合"): "zh",#按中英混合识别####不变
i18n("日英混合"): "ja",#按日英混合识别####不变
i18n("多语种混合"): "auto",#多语种启动切分识别语种
i18n("中文"): "all_zh", # 全部按中文识别
i18n("英文"): "en", # 全部按英文识别#######不变
i18n("日文"): "all_ja", # 全部按日文识别
i18n("中英混合"): "zh", # 按中英混合识别####不变
i18n("日英混合"): "ja", # 按日英混合识别####不变
i18n("多语种混合"): "auto", # 多语种启动切分识别语种
}
cut_method = {
i18n("不切"):"cut0",
i18n("不切"): "cut0",
i18n("凑四句一切"): "cut1",
i18n("凑50字一切"): "cut2",
i18n("按中文句号。切"): "cut3",
@ -80,23 +83,32 @@ if cnhubert_base_path is not None:
tts_config.cnhuhbert_base_path = cnhubert_base_path
if bert_path is not None:
tts_config.bert_base_path = bert_path
print(tts_config)
tts_pipline = TTS(tts_config)
gpt_path = tts_config.t2s_weights_path
sovits_path = tts_config.vits_weights_path
def inference(text, text_lang,
ref_audio_path, prompt_text,
prompt_lang, top_k,
top_p, temperature,
text_split_method, batch_size,
speed_factor, ref_text_free,
split_bucket,fragment_interval,
seed,
):
def inference(
text,
text_lang,
ref_audio_path,
prompt_text,
prompt_lang,
top_k,
top_p,
temperature,
text_split_method,
batch_size,
speed_factor,
ref_text_free,
split_bucket,
fragment_interval,
seed,
):
actual_seed = seed if seed not in [-1, "", None] else random.randrange(1 << 32)
inputs={
inputs = {
"text": text,
"text_lang": dict_language[text_lang],
"ref_audio_path": ref_audio_path,
@ -106,19 +118,20 @@ def inference(text, text_lang,
"top_p": top_p,
"temperature": temperature,
"text_split_method": cut_method[text_split_method],
"batch_size":int(batch_size),
"speed_factor":float(speed_factor),
"split_bucket":split_bucket,
"return_fragment":False,
"fragment_interval":fragment_interval,
"seed":actual_seed,
"batch_size": int(batch_size),
"speed_factor": float(speed_factor),
"split_bucket": split_bucket,
"return_fragment": False,
"fragment_interval": fragment_interval,
"seed": actual_seed,
}
for item in tts_pipline.run(inputs):
yield item, actual_seed
def custom_sort_key(s):
# 使用正则表达式提取字符串中的数字部分和非数字部分
parts = re.split('(\d+)', s)
parts = re.split("(\d+)", s)
# 将数字部分转换为整数，非数字部分保持不变
parts = [int(part) if part.isdigit() else part for part in parts]
return parts
@ -126,11 +139,16 @@ def custom_sort_key(s):
def change_choices():
SoVITS_names, GPT_names = get_weights_names()
return {"choices": sorted(SoVITS_names, key=custom_sort_key), "__type__": "update"}, {"choices": sorted(GPT_names, key=custom_sort_key), "__type__": "update"}
return {
"choices": sorted(SoVITS_names, key=custom_sort_key),
"__type__": "update",
}, {"choices": sorted(GPT_names, key=custom_sort_key), "__type__": "update"}
pretrained_sovits_name = "GPT_SoVITS/pretrained_models/s2G488k.pth"
pretrained_gpt_name = "GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt"
pretrained_gpt_name = (
"GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt"
)
SoVITS_weight_root = "SoVITS_weights"
GPT_weight_root = "GPT_weights"
os.makedirs(SoVITS_weight_root, exist_ok=True)
@ -140,122 +158,255 @@ os.makedirs(GPT_weight_root, exist_ok=True)
def get_weights_names():
SoVITS_names = [pretrained_sovits_name]
for name in os.listdir(SoVITS_weight_root):
if name.endswith(".pth"): SoVITS_names.append("%s/%s" % (SoVITS_weight_root, name))
if name.endswith(".pth"):
SoVITS_names.append("%s/%s" % (SoVITS_weight_root, name))
GPT_names = [pretrained_gpt_name]
for name in os.listdir(GPT_weight_root):
if name.endswith(".ckpt"): GPT_names.append("%s/%s" % (GPT_weight_root, name))
if name.endswith(".ckpt"):
GPT_names.append("%s/%s" % (GPT_weight_root, name))
return SoVITS_names, GPT_names
SoVITS_names, GPT_names = get_weights_names()
with gr.Blocks(title="GPT-SoVITS WebUI") as app:
gr.Markdown(
value=i18n("本软件以MIT协议开源, 作者不对软件具备任何控制力, 使用软件者、传播软件导出的声音者自负全责. <br>如不认可该条款, 则不能使用或引用软件包内任何代码和文件. 详见根目录<b>LICENSE</b>.")
)
with gr.Column():
# with gr.Group():
gr.Markdown(value=i18n("模型切换"))
with gr.Row():
GPT_dropdown = gr.Dropdown(label=i18n("GPT模型列表"), choices=sorted(GPT_names, key=custom_sort_key), value=gpt_path, interactive=True)
SoVITS_dropdown = gr.Dropdown(label=i18n("SoVITS模型列表"), choices=sorted(SoVITS_names, key=custom_sort_key), value=sovits_path, interactive=True)
refresh_button = gr.Button(i18n("刷新模型路径"), variant="primary")
refresh_button.click(fn=change_choices, inputs=[], outputs=[SoVITS_dropdown, GPT_dropdown])
SoVITS_dropdown.change(tts_pipline.init_vits_weights, [SoVITS_dropdown], [])
GPT_dropdown.change(tts_pipline.init_t2s_weights, [GPT_dropdown], [])
with gr.Row():
with gr.Column():
gr.Markdown(value=i18n("*请上传并填写参考信息"))
inp_ref = gr.Audio(label=i18n("请上传3~10秒内参考音频，超过会报错！"), type="filepath")
prompt_text = gr.Textbox(label=i18n("参考音频的文本"), value="", lines=2)
with gr.Row():
prompt_language = gr.Dropdown(
label=i18n("参考音频的语种"), choices=[i18n("中文"), i18n("英文"), i18n("日文"), i18n("中英混合"), i18n("日英混合"), i18n("多语种混合")], value=i18n("中文")
)
with gr.Column():
ref_text_free = gr.Checkbox(label=i18n("开启无参考文本模式。不填参考文本亦相当于开启。"), value=False, interactive=True, show_label=True)
gr.Markdown(i18n("使用无参考文本模式时建议使用微调的GPT，听不清参考音频说的啥(不晓得写啥)可以开，开启后无视填写的参考文本。"))
with gr.Column():
gr.Markdown(value=i18n("*请填写需要合成的目标文本和语种模式"))
text = gr.Textbox(label=i18n("需要合成的文本"), value="", lines=16, max_lines=16)
text_language = gr.Dropdown(
label=i18n("需要合成的语种"), choices=[i18n("中文"), i18n("英文"), i18n("日文"), i18n("中英混合"), i18n("日英混合"), i18n("多语种混合")], value=i18n("中文")
if __name__ == "__main__":
with gr.Blocks(title="GPT-SoVITS WebUI") as app:
gr.Markdown(
value=i18n(
"本软件以MIT协议开源, 作者不对软件具备任何控制力, 使用软件者、传播软件导出的声音者自负全责. <br>如不认可该条款, 则不能使用或引用软件包内任何代码和文件. 详见根目录<b>LICENSE</b>."
)
)
with gr.Group():
gr.Markdown(value=i18n("推理设置"))
with gr.Row():
with gr.Column():
batch_size = gr.Slider(minimum=1,maximum=200,step=1,label=i18n("batch_size"),value=20,interactive=True)
fragment_interval = gr.Slider(minimum=0.01,maximum=1,step=0.01,label=i18n("分段间隔(秒)"),value=0.3,interactive=True)
speed_factor = gr.Slider(minimum=0.25,maximum=4,step=0.05,label="speed_factor",value=1.0,interactive=True)
top_k = gr.Slider(minimum=1,maximum=100,step=1,label=i18n("top_k"),value=5,interactive=True)
top_p = gr.Slider(minimum=0,maximum=1,step=0.05,label=i18n("top_p"),value=1,interactive=True)
temperature = gr.Slider(minimum=0,maximum=1,step=0.05,label=i18n("temperature"),value=1,interactive=True)
with gr.Column():
how_to_cut = gr.Radio(
label=i18n("怎么切"),
choices=[i18n("不切"), i18n("凑四句一切"), i18n("凑50字一切"), i18n("按中文句号。切"), i18n("按英文句号.切"), i18n("按标点符号切"), ],
value=i18n("凑四句一切"),
with gr.Column():
# with gr.Group():
gr.Markdown(value=i18n("模型切换"))
with gr.Row():
GPT_dropdown = gr.Dropdown(
label=i18n("GPT模型列表"),
choices=sorted(GPT_names, key=custom_sort_key),
value=gpt_path,
interactive=True,
)
with gr.Row():
split_bucket = gr.Checkbox(label=i18n("数据分桶(可能会降低一点计算量，选就对了)"), value=True, interactive=True, show_label=True)
seed = gr.Number(label=i18n("随机种子"),value=-1)
# with gr.Column():
output = gr.Audio(label=i18n("输出的语音"))
with gr.Row():
inference_button = gr.Button(i18n("合成语音"), variant="primary")
stop_infer = gr.Button(i18n("终止合成"), variant="primary")
inference_button.click(
inference,
[
text,text_language, inp_ref,
prompt_text, prompt_language,
top_k, top_p, temperature,
how_to_cut, batch_size,
speed_factor, ref_text_free,
split_bucket,fragment_interval,
seed
],
[output, seed],
)
stop_infer.click(tts_pipline.stop, [], [])
SoVITS_dropdown = gr.Dropdown(
label=i18n("SoVITS模型列表"),
choices=sorted(SoVITS_names, key=custom_sort_key),
value=sovits_path,
interactive=True,
)
refresh_button = gr.Button(i18n("刷新模型路径"), variant="primary")
refresh_button.click(
fn=change_choices,
inputs=[],
outputs=[SoVITS_dropdown, GPT_dropdown],
)
SoVITS_dropdown.change(
tts_pipline.init_vits_weights, [SoVITS_dropdown], []
)
GPT_dropdown.change(tts_pipline.init_t2s_weights, [GPT_dropdown], [])
with gr.Group():
gr.Markdown(value=i18n("文本切分工具。太长的文本合成出来效果不一定好，所以太长建议先切。合成会根据文本的换行分开合成再拼起来。"))
with gr.Row():
text_inp = gr.Textbox(label=i18n("需要合成的切分前文本"), value="", lines=4)
with gr.Column():
_how_to_cut = gr.Radio(
label=i18n("怎么切"),
choices=[i18n("不切"), i18n("凑四句一切"), i18n("凑50字一切"), i18n("按中文句号。切"), i18n("按英文句号.切"), i18n("按标点符号切"), ],
value=i18n("凑四句一切"),
gr.Markdown(value=i18n("*请上传并填写参考信息"))
inp_ref = gr.Audio(
label=i18n("请上传3~10秒内参考音频，超过会报错！"), type="filepath"
)
prompt_text = gr.Textbox(
label=i18n("参考音频的文本"), value="", lines=2
)
with gr.Row():
prompt_language = gr.Dropdown(
label=i18n("参考音频的语种"),
choices=[
i18n("中文"),
i18n("英文"),
i18n("日文"),
i18n("中英混合"),
i18n("日英混合"),
i18n("多语种混合"),
],
value=i18n("中文"),
)
with gr.Column():
ref_text_free = gr.Checkbox(
label=i18n(
"开启无参考文本模式。不填参考文本亦相当于开启。"
),
value=False,
interactive=True,
show_label=True,
)
gr.Markdown(
i18n(
"使用无参考文本模式时建议使用微调的GPT，听不清参考音频说的啥(不晓得写啥)可以开，开启后无视填写的参考文本。"
)
)
cut_text= gr.Button(i18n("切分"), variant="primary")
def to_cut(text_inp, how_to_cut):
if len(text_inp.strip()) == 0 or text_inp==[]:
return ""
method = get_method(cut_method[how_to_cut])
return method(text_inp)
text_opt = gr.Textbox(label=i18n("切分后文本"), value="", lines=4)
cut_text.click(to_cut, [text_inp, _how_to_cut], [text_opt])
gr.Markdown(value=i18n("后续将支持转音素、手工修改音素、语音合成分步执行。"))
app.queue(concurrency_count=511, max_size=1022).launch(
server_name="0.0.0.0",
inbrowser=True,
share=is_share,
server_port=infer_ttswebui,
quiet=True,
)
with gr.Column():
gr.Markdown(value=i18n("*请填写需要合成的目标文本和语种模式"))
text = gr.Textbox(
label=i18n("需要合成的文本"), value="", lines=16, max_lines=16
)
text_language = gr.Dropdown(
label=i18n("需要合成的语种"),
choices=[
i18n("中文"),
i18n("英文"),
i18n("日文"),
i18n("中英混合"),
i18n("日英混合"),
i18n("多语种混合"),
],
value=i18n("中文"),
)
with gr.Group():
gr.Markdown(value=i18n("推理设置"))
with gr.Row():
with gr.Column():
batch_size = gr.Slider(
minimum=1,
maximum=200,
step=1,
label=i18n("batch_size"),
value=20,
interactive=True,
)
fragment_interval = gr.Slider(
minimum=0.01,
maximum=1,
step=0.01,
label=i18n("分段间隔(秒)"),
value=0.3,
interactive=True,
)
speed_factor = gr.Slider(
minimum=0.25,
maximum=4,
step=0.05,
label="speed_factor",
value=1.0,
interactive=True,
)
top_k = gr.Slider(
minimum=1,
maximum=100,
step=1,
label=i18n("top_k"),
value=5,
interactive=True,
)
top_p = gr.Slider(
minimum=0,
maximum=1,
step=0.05,
label=i18n("top_p"),
value=1,
interactive=True,
)
temperature = gr.Slider(
minimum=0,
maximum=1,
step=0.05,
label=i18n("temperature"),
value=1,
interactive=True,
)
with gr.Column():
how_to_cut = gr.Radio(
label=i18n("怎么切"),
choices=[
i18n("不切"),
i18n("凑四句一切"),
i18n("凑50字一切"),
i18n("按中文句号。切"),
i18n("按英文句号.切"),
i18n("按标点符号切"),
],
value=i18n("凑四句一切"),
interactive=True,
)
with gr.Row():
split_bucket = gr.Checkbox(
label=i18n("数据分桶(可能会降低一点计算量，选就对了)"),
value=True,
interactive=True,
show_label=True,
)
seed = gr.Number(label=i18n("随机种子"), value=-1)
# with gr.Column():
output = gr.Audio(label=i18n("输出的语音"))
with gr.Row():
inference_button = gr.Button(
i18n("合成语音"), variant="primary"
)
stop_infer = gr.Button(i18n("终止合成"), variant="primary")
inference_button.click(
inference,
[
text,
text_language,
inp_ref,
prompt_text,
prompt_language,
top_k,
top_p,
temperature,
how_to_cut,
batch_size,
speed_factor,
ref_text_free,
split_bucket,
fragment_interval,
seed,
],
[output, seed],
)
stop_infer.click(tts_pipline.stop, [], [])
with gr.Group():
gr.Markdown(
value=i18n(
"文本切分工具。太长的文本合成出来效果不一定好，所以太长建议先切。合成会根据文本的换行分开合成再拼起来。"
)
)
with gr.Row():
text_inp = gr.Textbox(
label=i18n("需要合成的切分前文本"), value="", lines=4
)
with gr.Column():
_how_to_cut = gr.Radio(
label=i18n("怎么切"),
choices=[
i18n("不切"),
i18n("凑四句一切"),
i18n("凑50字一切"),
i18n("按中文句号。切"),
i18n("按英文句号.切"),
i18n("按标点符号切"),
],
value=i18n("凑四句一切"),
interactive=True,
)
cut_text = gr.Button(i18n("切分"), variant="primary")
def to_cut(text_inp, how_to_cut):
if len(text_inp.strip()) == 0 or text_inp == []:
return ""
method = get_method(cut_method[how_to_cut])
return method(text_inp)
text_opt = gr.Textbox(label=i18n("切分后文本"), value="", lines=4)
cut_text.click(to_cut, [text_inp, _how_to_cut], [text_opt])
gr.Markdown(
value=i18n("后续将支持转音素、手工修改音素、语音合成分步执行。")
)
app.queue(concurrency_count=511, max_size=1022).launch(
server_name="0.0.0.0",
inbrowser=True,
share=is_share,
server_port=infer_ttswebui,
quiet=True,
)