mirror of
https://github.com/RVC-Boss/GPT-SoVITS.git
synced 2025-08-18 15:59:51 +08:00
turn it into a package
This commit is contained in:
Jarod Mica
parent
a70e1ad30c
commit
71b2fe69ea
@ -1,60 +1,66 @@
|
||||
from copy import deepcopy
|
||||
import math
|
||||
import os, sys, gc
|
||||
import random
|
||||
import traceback
|
||||
|
||||
from tqdm import tqdm
|
||||
|
||||
now_dir = os.getcwd()
|
||||
sys.path.append(now_dir)
|
||||
import ffmpeg
|
||||
import os
|
||||
from typing import Generator, List, Tuple, Union
|
||||
import librosa
|
||||
import numpy as np
|
||||
import random
|
||||
import torch
|
||||
import torch.nn.functional as F
|
||||
import traceback
|
||||
import yaml
|
||||
from transformers import AutoModelForMaskedLM, AutoTokenizer
|
||||
|
||||
from huggingface_hub import snapshot_download, hf_hub_download
|
||||
from importlib.resources import files
|
||||
from time import time as ttime
|
||||
from typing import Generator, List, Tuple, Union
|
||||
from tqdm import tqdm
|
||||
|
||||
|
||||
from AR.models.t2s_lightning_module import Text2SemanticLightningModule
|
||||
from feature_extractor.cnhubert import CNHubert
|
||||
from module.models import SynthesizerTrn
|
||||
import librosa
|
||||
from time import time as ttime
|
||||
from module.mel_processing import spectrogram_torch
|
||||
from tools.i18n.i18n import I18nAuto, scan_language_list
|
||||
from tools.my_utils import load_audio
|
||||
from module.mel_processing import spectrogram_torch
|
||||
from transformers import AutoModelForMaskedLM, AutoTokenizer
|
||||
from TTS_infer_pack.text_segmentation_method import splits
|
||||
from TTS_infer_pack.TextPreprocessor import TextPreprocessor
|
||||
|
||||
language=os.environ.get("language","Auto")
|
||||
language=sys.argv[-1] if sys.argv[-1] in scan_language_list() else language
|
||||
i18n = I18nAuto(language=language)
|
||||
LIBRARY_NAME = "GPT_SoVITS"
|
||||
|
||||
# configs/tts_infer.yaml
|
||||
"""
|
||||
custom:
|
||||
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
|
||||
bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: pretrained_models/chinese-hubert-base
|
||||
device: cpu
|
||||
is_half: false
|
||||
t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
version: v2
|
||||
default:
|
||||
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
|
||||
bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: pretrained_models/chinese-hubert-base
|
||||
device: cpu
|
||||
is_half: false
|
||||
t2s_weights_path: GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
|
||||
vits_weights_path: GPT_SoVITS/pretrained_models/s2G488k.pth
|
||||
t2s_weights_path: pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
|
||||
vits_weights_path: pretrained_models/s2G488k.pth
|
||||
version: v1
|
||||
default_v2:
|
||||
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
|
||||
bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: pretrained_models/chinese-hubert-base
|
||||
device: cpu
|
||||
is_half: false
|
||||
t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
version: v2
|
||||
"""
|
||||
|
||||
@ -86,19 +92,19 @@ class TTS_Config:
|
||||
"device": "cpu",
|
||||
"is_half": False,
|
||||
"version": "v1",
|
||||
"t2s_weights_path": "GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt",
|
||||
"vits_weights_path": "GPT_SoVITS/pretrained_models/s2G488k.pth",
|
||||
"cnhuhbert_base_path": "GPT_SoVITS/pretrained_models/chinese-hubert-base",
|
||||
"bert_base_path": "GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large",
|
||||
"t2s_weights_path": "pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt",
|
||||
"vits_weights_path": "pretrained_models/s2G488k.pth",
|
||||
"cnhuhbert_base_path": "pretrained_models/chinese-hubert-base",
|
||||
"bert_base_path": "pretrained_models/chinese-roberta-wwm-ext-large",
|
||||
},
|
||||
"default_v2":{
|
||||
"device": "cpu",
|
||||
"is_half": False,
|
||||
"version": "v2",
|
||||
"t2s_weights_path": "GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt",
|
||||
"vits_weights_path": "GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth",
|
||||
"cnhuhbert_base_path": "GPT_SoVITS/pretrained_models/chinese-hubert-base",
|
||||
"bert_base_path": "GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large",
|
||||
"t2s_weights_path": "pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt",
|
||||
"vits_weights_path": "pretrained_models/gsv-v2final-pretrained/s2G2333k.pth",
|
||||
"cnhuhbert_base_path": "pretrained_models/chinese-hubert-base",
|
||||
"bert_base_path": "pretrained_models/chinese-roberta-wwm-ext-large",
|
||||
},
|
||||
}
|
||||
configs:dict = None
|
||||
@ -120,7 +126,7 @@ class TTS_Config:
|
||||
def __init__(self, configs: Union[dict, str]=None):
|
||||
|
||||
# 设置默认配置文件路径
|
||||
configs_base_path:str = "GPT_SoVITS/configs/"
|
||||
configs_base_path:str = "configs/"
|
||||
os.makedirs(configs_base_path, exist_ok=True)
|
||||
self.configs_path:str = os.path.join(configs_base_path, "tts_infer.yaml")
|
||||
|
||||
@ -152,23 +158,24 @@ class TTS_Config:
|
||||
self.bert_base_path = self.configs.get("bert_base_path", None)
|
||||
self.cnhuhbert_base_path = self.configs.get("cnhuhbert_base_path", None)
|
||||
self.languages = self.v2_languages if self.version=="v2" else self.v1_languages
|
||||
|
||||
|
||||
if (self.t2s_weights_path in [None, ""]) or (not os.path.exists(self.t2s_weights_path)):
|
||||
self.t2s_weights_path = self.default_configs[default_config_key]['t2s_weights_path']
|
||||
self.t2s_weights_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['t2s_weights_path']))
|
||||
print(f"fall back to default t2s_weights_path: {self.t2s_weights_path}")
|
||||
if (self.vits_weights_path in [None, ""]) or (not os.path.exists(self.vits_weights_path)):
|
||||
self.vits_weights_path = self.default_configs[default_config_key]['vits_weights_path']
|
||||
self.vits_weights_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['vits_weights_path']))
|
||||
print(f"fall back to default vits_weights_path: {self.vits_weights_path}")
|
||||
if (self.bert_base_path in [None, ""]) or (not os.path.exists(self.bert_base_path)):
|
||||
self.bert_base_path = self.default_configs[default_config_key]['bert_base_path']
|
||||
self.bert_base_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['bert_base_path']))
|
||||
print(f"fall back to default bert_base_path: {self.bert_base_path}")
|
||||
if (self.cnhuhbert_base_path in [None, ""]) or (not os.path.exists(self.cnhuhbert_base_path)):
|
||||
self.cnhuhbert_base_path = self.default_configs[default_config_key]['cnhuhbert_base_path']
|
||||
self.cnhuhbert_base_path = str(files(LIBRARY_NAME).joinpath(self.default_configs[default_config_key]['cnhuhbert_base_path']))
|
||||
print(f"fall back to default cnhuhbert_base_path: {self.cnhuhbert_base_path}")
|
||||
|
||||
repo_name="lj1995/GPT-SoVITS"
|
||||
snapshot_download(repo_id=repo_name, local_dir=os.path.dirname(self.bert_base_path))
|
||||
self.update_configs()
|
||||
|
||||
|
||||
self.max_sec = None
|
||||
self.hz:int = 50
|
||||
self.semantic_frame_rate:str = "25hz"
|
||||
|
||||
@ -114,6 +114,8 @@ from fastapi import FastAPI, Request, HTTPException, Response
|
||||
from fastapi.responses import StreamingResponse, JSONResponse
|
||||
from fastapi import FastAPI, UploadFile, File
|
||||
import uvicorn
|
||||
|
||||
from importlib.resources import files
|
||||
from io import BytesIO
|
||||
from tools.i18n.i18n import I18nAuto
|
||||
from GPT_SoVITS.TTS_infer_pack.TTS import TTS, TTS_Config
|
||||
@ -125,7 +127,7 @@ i18n = I18nAuto()
|
||||
cut_method_names = get_cut_method_names()
|
||||
|
||||
parser = argparse.ArgumentParser(description="GPT-SoVITS api")
|
||||
parser.add_argument("-c", "--tts_config", type=str, default="GPT_SoVITS/configs/tts_infer.yaml", help="tts_infer路径")
|
||||
parser.add_argument("-c", "--tts_config", type=str, default=None, help="tts_infer路径")
|
||||
parser.add_argument("-a", "--bind_addr", type=str, default="127.0.0.1", help="default: 127.0.0.1")
|
||||
parser.add_argument("-p", "--port", type=int, default="9880", help="default: 9880")
|
||||
args = parser.parse_args()
|
||||
@ -136,7 +138,7 @@ host = args.bind_addr
|
||||
argv = sys.argv
|
||||
|
||||
if config_path in [None, ""]:
|
||||
config_path = "GPT-SoVITS/configs/tts_infer.yaml"
|
||||
config_path = str(files("GPT_SoVITS").joinpath("configs/tts_infer.yaml"))
|
||||
|
||||
tts_config = TTS_Config(config_path)
|
||||
print(tts_config)
|
||||
@ -394,7 +396,7 @@ async def tts_get_endpoint(
|
||||
|
||||
@APP.post("/tts")
|
||||
async def tts_post_endpoint(request: TTS_Request):
|
||||
req = request.dict()
|
||||
req = request.model_dump()
|
||||
return await tts_handle(req)
|
||||
|
||||
|
||||
@ -449,7 +451,8 @@ async def set_sovits_weights(weights_path: str = None):
|
||||
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
def main():
|
||||
global port, host, argv
|
||||
try:
|
||||
if host == 'None': # 在调用时使用 -a None 参数，可以让api监听双栈
|
||||
host = None
|
||||
@ -458,3 +461,6 @@ if __name__ == "__main__":
|
||||
traceback.print_exc()
|
||||
os.kill(os.getpid(), signal.SIGTERM)
|
||||
exit(0)
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
@ -1,24 +1,24 @@
|
||||
custom:
|
||||
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
|
||||
bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: pretrained_models/chinese-hubert-base
|
||||
device: cuda
|
||||
is_half: true
|
||||
t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
version: v2
|
||||
vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
default:
|
||||
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
|
||||
bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: pretrained_models/chinese-hubert-base
|
||||
device: cpu
|
||||
is_half: false
|
||||
t2s_weights_path: GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
|
||||
t2s_weights_path: pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt
|
||||
version: v1
|
||||
vits_weights_path: GPT_SoVITS/pretrained_models/s2G488k.pth
|
||||
vits_weights_path: pretrained_models/s2G488k.pth
|
||||
default_v2:
|
||||
bert_base_path: GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: GPT_SoVITS/pretrained_models/chinese-hubert-base
|
||||
bert_base_path: pretrained_models/chinese-roberta-wwm-ext-large
|
||||
cnhuhbert_base_path: pretrained_models/chinese-hubert-base
|
||||
device: cpu
|
||||
is_half: false
|
||||
t2s_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
t2s_weights_path: pretrained_models/gsv-v2final-pretrained/s1bert25hz-5kh-longer-epoch=12-step=369668.ckpt
|
||||
version: v2
|
||||
vits_weights_path: GPT_SoVITS/pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
vits_weights_path: pretrained_models/gsv-v2final-pretrained/s2G2333k.pth
|
||||
|
||||
@ -1,33 +1,33 @@
|
||||
import os,argparse
|
||||
import traceback
|
||||
|
||||
from modelscope.pipelines import pipeline
|
||||
from modelscope.utils.constant import Tasks
|
||||
from tqdm import tqdm
|
||||
|
||||
path_denoise = 'tools/denoise-model/speech_frcrn_ans_cirm_16k'
|
||||
path_denoise = path_denoise if os.path.exists(path_denoise) else "damo/speech_frcrn_ans_cirm_16k"
|
||||
ans = pipeline(Tasks.acoustic_noise_suppression,model=path_denoise)
|
||||
def execute_denoise(input_folder,output_folder):
|
||||
os.makedirs(output_folder,exist_ok=True)
|
||||
# print(input_folder)
|
||||
# print(list(os.listdir(input_folder).sort()))
|
||||
for name in tqdm(os.listdir(input_folder)):
|
||||
try:
|
||||
ans("%s/%s"%(input_folder,name),output_path='%s/%s'%(output_folder,name))
|
||||
except:
|
||||
traceback.print_exc()
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument("-i", "--input_folder", type=str, required=True,
|
||||
help="Path to the folder containing WAV files.")
|
||||
parser.add_argument("-o", "--output_folder", type=str, required=True,
|
||||
help="Output folder to store transcriptions.")
|
||||
parser.add_argument("-p", "--precision", type=str, default='float16', choices=['float16','float32'],
|
||||
help="fp16 or fp32")#还没接入
|
||||
cmd = parser.parse_args()
|
||||
execute_denoise(
|
||||
input_folder = cmd.input_folder,
|
||||
output_folder = cmd.output_folder,
|
||||
import os,argparse
|
||||
import traceback
|
||||
|
||||
from modelscope.pipelines import pipeline
|
||||
from modelscope.utils.constant import Tasks
|
||||
from tqdm import tqdm
|
||||
|
||||
path_denoise = 'tools/denoise-model/speech_frcrn_ans_cirm_16k'
|
||||
path_denoise = path_denoise if os.path.exists(path_denoise) else "damo/speech_frcrn_ans_cirm_16k"
|
||||
ans = pipeline(Tasks.acoustic_noise_suppression,model=path_denoise)
|
||||
def execute_denoise(input_folder,output_folder):
|
||||
os.makedirs(output_folder,exist_ok=True)
|
||||
# print(input_folder)
|
||||
# print(list(os.listdir(input_folder).sort()))
|
||||
for name in tqdm(os.listdir(input_folder)):
|
||||
try:
|
||||
ans("%s/%s"%(input_folder,name),output_path='%s/%s'%(output_folder,name))
|
||||
except:
|
||||
traceback.print_exc()
|
||||
|
||||
if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser()
|
||||
parser.add_argument("-i", "--input_folder", type=str, required=True,
|
||||
help="Path to the folder containing WAV files.")
|
||||
parser.add_argument("-o", "--output_folder", type=str, required=True,
|
||||
help="Output folder to store transcriptions.")
|
||||
parser.add_argument("-p", "--precision", type=str, default='float16', choices=['float16','float32'],
|
||||
help="fp16 or fp32")#还没接入
|
||||
cmd = parser.parse_args()
|
||||
execute_denoise(
|
||||
input_folder = cmd.input_folder,
|
||||
output_folder = cmd.output_folder,
|
||||
)
|
||||
@ -1,115 +1,115 @@
|
||||
import platform,os,traceback
|
||||
import ffmpeg
|
||||
import numpy as np
|
||||
import gradio as gr
|
||||
from tools.i18n.i18n import I18nAuto
|
||||
import pandas as pd
|
||||
i18n = I18nAuto(language=os.environ.get('language','Auto'))
|
||||
|
||||
def load_audio(file, sr):
|
||||
try:
|
||||
# https://github.com/openai/whisper/blob/main/whisper/audio.py#L26
|
||||
# This launches a subprocess to decode audio while down-mixing and resampling as necessary.
|
||||
# Requires the ffmpeg CLI and `ffmpeg-python` package to be installed.
|
||||
file = clean_path(file) # 防止小白拷路径头尾带了空格和"和回车
|
||||
if os.path.exists(file) == False:
|
||||
raise RuntimeError(
|
||||
"You input a wrong audio path that does not exists, please fix it!"
|
||||
)
|
||||
out, _ = (
|
||||
ffmpeg.input(file, threads=0)
|
||||
.output("-", format="f32le", acodec="pcm_f32le", ac=1, ar=sr)
|
||||
.run(cmd=["ffmpeg", "-nostdin"], capture_stdout=True, capture_stderr=True)
|
||||
)
|
||||
except Exception as e:
|
||||
traceback.print_exc()
|
||||
raise RuntimeError(i18n("音频加载失败"))
|
||||
|
||||
return np.frombuffer(out, np.float32).flatten()
|
||||
|
||||
|
||||
def clean_path(path_str:str):
|
||||
if path_str.endswith(('\\','/')):
|
||||
return clean_path(path_str[0:-1])
|
||||
path_str = path_str.replace('/', os.sep).replace('\\', os.sep)
|
||||
return path_str.strip(" ").strip('\'').strip("\n").strip('"').strip(" ").strip("\u202a")
|
||||
|
||||
|
||||
def check_for_existance(file_list:list=None,is_train=False,is_dataset_processing=False):
|
||||
files_status=[]
|
||||
if is_train == True and file_list:
|
||||
file_list.append(os.path.join(file_list[0],'2-name2text.txt'))
|
||||
file_list.append(os.path.join(file_list[0],'3-bert'))
|
||||
file_list.append(os.path.join(file_list[0],'4-cnhubert'))
|
||||
file_list.append(os.path.join(file_list[0],'5-wav32k'))
|
||||
file_list.append(os.path.join(file_list[0],'6-name2semantic.tsv'))
|
||||
for file in file_list:
|
||||
if os.path.exists(file):files_status.append(True)
|
||||
else:files_status.append(False)
|
||||
if sum(files_status)!=len(files_status):
|
||||
if is_train:
|
||||
for file,status in zip(file_list,files_status):
|
||||
if status:pass
|
||||
else:gr.Warning(file)
|
||||
gr.Warning(i18n('以下文件或文件夹不存在'))
|
||||
return False
|
||||
elif is_dataset_processing:
|
||||
if files_status[0]:
|
||||
return True
|
||||
elif not files_status[0]:
|
||||
gr.Warning(file_list[0])
|
||||
elif not files_status[1] and file_list[1]:
|
||||
gr.Warning(file_list[1])
|
||||
gr.Warning(i18n('以下文件或文件夹不存在'))
|
||||
return False
|
||||
else:
|
||||
if file_list[0]:
|
||||
gr.Warning(file_list[0])
|
||||
gr.Warning(i18n('以下文件或文件夹不存在'))
|
||||
else:
|
||||
gr.Warning(i18n('路径不能为空'))
|
||||
return False
|
||||
return True
|
||||
|
||||
def check_details(path_list=None,is_train=False,is_dataset_processing=False):
|
||||
if is_dataset_processing:
|
||||
list_path, audio_path = path_list
|
||||
if (not list_path.endswith('.list')):
|
||||
gr.Warning(i18n('请填入正确的List路径'))
|
||||
return
|
||||
if audio_path:
|
||||
if not os.path.isdir(audio_path):
|
||||
gr.Warning(i18n('请填入正确的音频文件夹路径'))
|
||||
return
|
||||
with open(list_path,"r",encoding="utf8")as f:
|
||||
line=f.readline().strip("\n").split("\n")
|
||||
wav_name, _, __, ___ = line[0].split("|")
|
||||
wav_name=clean_path(wav_name)
|
||||
if (audio_path != "" and audio_path != None):
|
||||
wav_name = os.path.basename(wav_name)
|
||||
wav_path = "%s/%s"%(audio_path, wav_name)
|
||||
else:
|
||||
wav_path=wav_name
|
||||
if os.path.exists(wav_path):
|
||||
...
|
||||
else:
|
||||
gr.Warning(i18n('路径错误'))
|
||||
return
|
||||
if is_train:
|
||||
path_list.append(os.path.join(path_list[0],'2-name2text.txt'))
|
||||
path_list.append(os.path.join(path_list[0],'4-cnhubert'))
|
||||
path_list.append(os.path.join(path_list[0],'5-wav32k'))
|
||||
path_list.append(os.path.join(path_list[0],'6-name2semantic.tsv'))
|
||||
phone_path, hubert_path, wav_path, semantic_path = path_list[1:]
|
||||
with open(phone_path,'r',encoding='utf-8') as f:
|
||||
if f.read(1):...
|
||||
else:gr.Warning(i18n('缺少音素数据集'))
|
||||
if os.listdir(hubert_path):...
|
||||
else:gr.Warning(i18n('缺少Hubert数据集'))
|
||||
if os.listdir(wav_path):...
|
||||
else:gr.Warning(i18n('缺少音频数据集'))
|
||||
df = pd.read_csv(
|
||||
semantic_path, delimiter="\t", encoding="utf-8"
|
||||
)
|
||||
if len(df) >= 1:...
|
||||
else:gr.Warning(i18n('缺少语义数据集'))
|
||||
import platform,os,traceback
|
||||
import ffmpeg
|
||||
import numpy as np
|
||||
import gradio as gr
|
||||
from tools.i18n.i18n import I18nAuto
|
||||
import pandas as pd
|
||||
i18n = I18nAuto(language=os.environ.get('language','Auto'))
|
||||
|
||||
def load_audio(file, sr):
|
||||
try:
|
||||
# https://github.com/openai/whisper/blob/main/whisper/audio.py#L26
|
||||
# This launches a subprocess to decode audio while down-mixing and resampling as necessary.
|
||||
# Requires the ffmpeg CLI and `ffmpeg-python` package to be installed.
|
||||
file = clean_path(file) # 防止小白拷路径头尾带了空格和"和回车
|
||||
if os.path.exists(file) == False:
|
||||
raise RuntimeError(
|
||||
"You input a wrong audio path that does not exists, please fix it!"
|
||||
)
|
||||
out, _ = (
|
||||
ffmpeg.input(file, threads=0)
|
||||
.output("-", format="f32le", acodec="pcm_f32le", ac=1, ar=sr)
|
||||
.run(cmd=["ffmpeg", "-nostdin"], capture_stdout=True, capture_stderr=True)
|
||||
)
|
||||
except Exception as e:
|
||||
traceback.print_exc()
|
||||
raise RuntimeError(i18n("音频加载失败"))
|
||||
|
||||
return np.frombuffer(out, np.float32).flatten()
|
||||
|
||||
|
||||
def clean_path(path_str:str):
|
||||
if path_str.endswith(('\\','/')):
|
||||
return clean_path(path_str[0:-1])
|
||||
path_str = path_str.replace('/', os.sep).replace('\\', os.sep)
|
||||
return path_str.strip(" ").strip('\'').strip("\n").strip('"').strip(" ").strip("\u202a")
|
||||
|
||||
|
||||
def check_for_existance(file_list:list=None,is_train=False,is_dataset_processing=False):
|
||||
files_status=[]
|
||||
if is_train == True and file_list:
|
||||
file_list.append(os.path.join(file_list[0],'2-name2text.txt'))
|
||||
file_list.append(os.path.join(file_list[0],'3-bert'))
|
||||
file_list.append(os.path.join(file_list[0],'4-cnhubert'))
|
||||
file_list.append(os.path.join(file_list[0],'5-wav32k'))
|
||||
file_list.append(os.path.join(file_list[0],'6-name2semantic.tsv'))
|
||||
for file in file_list:
|
||||
if os.path.exists(file):files_status.append(True)
|
||||
else:files_status.append(False)
|
||||
if sum(files_status)!=len(files_status):
|
||||
if is_train:
|
||||
for file,status in zip(file_list,files_status):
|
||||
if status:pass
|
||||
else:gr.Warning(file)
|
||||
gr.Warning(i18n('以下文件或文件夹不存在'))
|
||||
return False
|
||||
elif is_dataset_processing:
|
||||
if files_status[0]:
|
||||
return True
|
||||
elif not files_status[0]:
|
||||
gr.Warning(file_list[0])
|
||||
elif not files_status[1] and file_list[1]:
|
||||
gr.Warning(file_list[1])
|
||||
gr.Warning(i18n('以下文件或文件夹不存在'))
|
||||
return False
|
||||
else:
|
||||
if file_list[0]:
|
||||
gr.Warning(file_list[0])
|
||||
gr.Warning(i18n('以下文件或文件夹不存在'))
|
||||
else:
|
||||
gr.Warning(i18n('路径不能为空'))
|
||||
return False
|
||||
return True
|
||||
|
||||
def check_details(path_list=None,is_train=False,is_dataset_processing=False):
|
||||
if is_dataset_processing:
|
||||
list_path, audio_path = path_list
|
||||
if (not list_path.endswith('.list')):
|
||||
gr.Warning(i18n('请填入正确的List路径'))
|
||||
return
|
||||
if audio_path:
|
||||
if not os.path.isdir(audio_path):
|
||||
gr.Warning(i18n('请填入正确的音频文件夹路径'))
|
||||
return
|
||||
with open(list_path,"r",encoding="utf8")as f:
|
||||
line=f.readline().strip("\n").split("\n")
|
||||
wav_name, _, __, ___ = line[0].split("|")
|
||||
wav_name=clean_path(wav_name)
|
||||
if (audio_path != "" and audio_path != None):
|
||||
wav_name = os.path.basename(wav_name)
|
||||
wav_path = "%s/%s"%(audio_path, wav_name)
|
||||
else:
|
||||
wav_path=wav_name
|
||||
if os.path.exists(wav_path):
|
||||
...
|
||||
else:
|
||||
gr.Warning(i18n('路径错误'))
|
||||
return
|
||||
if is_train:
|
||||
path_list.append(os.path.join(path_list[0],'2-name2text.txt'))
|
||||
path_list.append(os.path.join(path_list[0],'4-cnhubert'))
|
||||
path_list.append(os.path.join(path_list[0],'5-wav32k'))
|
||||
path_list.append(os.path.join(path_list[0],'6-name2semantic.tsv'))
|
||||
phone_path, hubert_path, wav_path, semantic_path = path_list[1:]
|
||||
with open(phone_path,'r',encoding='utf-8') as f:
|
||||
if f.read(1):...
|
||||
else:gr.Warning(i18n('缺少音素数据集'))
|
||||
if os.listdir(hubert_path):...
|
||||
else:gr.Warning(i18n('缺少Hubert数据集'))
|
||||
if os.listdir(wav_path):...
|
||||
else:gr.Warning(i18n('缺少音频数据集'))
|
||||
df = pd.read_csv(
|
||||
semantic_path, delimiter="\t", encoding="utf-8"
|
||||
)
|
||||
if len(df) >= 1:...
|
||||
else:gr.Warning(i18n('缺少语义数据集'))
|
||||
@ -1,48 +1,48 @@
|
||||
import os,sys,numpy as np
|
||||
import traceback
|
||||
from scipy.io import wavfile
|
||||
# parent_directory = os.path.dirname(os.path.abspath(__file__))
|
||||
# sys.path.append(parent_directory)
|
||||
from tools.my_utils import load_audio
|
||||
from slicer2 import Slicer
|
||||
|
||||
def slice(inp,opt_root,threshold,min_length,min_interval,hop_size,max_sil_kept,_max,alpha,i_part,all_part):
|
||||
os.makedirs(opt_root,exist_ok=True)
|
||||
if os.path.isfile(inp):
|
||||
input=[inp]
|
||||
elif os.path.isdir(inp):
|
||||
input=[os.path.join(inp, name) for name in sorted(list(os.listdir(inp)))]
|
||||
else:
|
||||
return "输入路径存在但既不是文件也不是文件夹"
|
||||
slicer = Slicer(
|
||||
sr=32000, # 长音频采样率
|
||||
threshold= int(threshold), # 音量小于这个值视作静音的备选切割点
|
||||
min_length= int(min_length), # 每段最小多长，如果第一段太短一直和后面段连起来直到超过这个值
|
||||
min_interval= int(min_interval), # 最短切割间隔
|
||||
hop_size= int(hop_size), # 怎么算音量曲线，越小精度越大计算量越高（不是精度越大效果越好）
|
||||
max_sil_kept= int(max_sil_kept), # 切完后静音最多留多长
|
||||
)
|
||||
_max=float(_max)
|
||||
alpha=float(alpha)
|
||||
for inp_path in input[int(i_part)::int(all_part)]:
|
||||
# print(inp_path)
|
||||
try:
|
||||
name = os.path.basename(inp_path)
|
||||
audio = load_audio(inp_path, 32000)
|
||||
# print(audio.shape)
|
||||
for chunk, start, end in slicer.slice(audio): # start和end是帧数
|
||||
tmp_max = np.abs(chunk).max()
|
||||
if(tmp_max>1):chunk/=tmp_max
|
||||
chunk = (chunk / tmp_max * (_max * alpha)) + (1 - alpha) * chunk
|
||||
wavfile.write(
|
||||
"%s/%s_%010d_%010d.wav" % (opt_root, name, start, end),
|
||||
32000,
|
||||
# chunk.astype(np.float32),
|
||||
(chunk * 32767).astype(np.int16),
|
||||
)
|
||||
except:
|
||||
print(inp_path,"->fail->",traceback.format_exc())
|
||||
return "执行完毕，请检查输出文件"
|
||||
|
||||
print(slice(*sys.argv[1:]))
|
||||
|
||||
import os,sys,numpy as np
|
||||
import traceback
|
||||
from scipy.io import wavfile
|
||||
# parent_directory = os.path.dirname(os.path.abspath(__file__))
|
||||
# sys.path.append(parent_directory)
|
||||
from tools.my_utils import load_audio
|
||||
from slicer2 import Slicer
|
||||
|
||||
def slice(inp,opt_root,threshold,min_length,min_interval,hop_size,max_sil_kept,_max,alpha,i_part,all_part):
|
||||
os.makedirs(opt_root,exist_ok=True)
|
||||
if os.path.isfile(inp):
|
||||
input=[inp]
|
||||
elif os.path.isdir(inp):
|
||||
input=[os.path.join(inp, name) for name in sorted(list(os.listdir(inp)))]
|
||||
else:
|
||||
return "输入路径存在但既不是文件也不是文件夹"
|
||||
slicer = Slicer(
|
||||
sr=32000, # 长音频采样率
|
||||
threshold= int(threshold), # 音量小于这个值视作静音的备选切割点
|
||||
min_length= int(min_length), # 每段最小多长，如果第一段太短一直和后面段连起来直到超过这个值
|
||||
min_interval= int(min_interval), # 最短切割间隔
|
||||
hop_size= int(hop_size), # 怎么算音量曲线，越小精度越大计算量越高（不是精度越大效果越好）
|
||||
max_sil_kept= int(max_sil_kept), # 切完后静音最多留多长
|
||||
)
|
||||
_max=float(_max)
|
||||
alpha=float(alpha)
|
||||
for inp_path in input[int(i_part)::int(all_part)]:
|
||||
# print(inp_path)
|
||||
try:
|
||||
name = os.path.basename(inp_path)
|
||||
audio = load_audio(inp_path, 32000)
|
||||
# print(audio.shape)
|
||||
for chunk, start, end in slicer.slice(audio): # start和end是帧数
|
||||
tmp_max = np.abs(chunk).max()
|
||||
if(tmp_max>1):chunk/=tmp_max
|
||||
chunk = (chunk / tmp_max * (_max * alpha)) + (1 - alpha) * chunk
|
||||
wavfile.write(
|
||||
"%s/%s_%010d_%010d.wav" % (opt_root, name, start, end),
|
||||
32000,
|
||||
# chunk.astype(np.float32),
|
||||
(chunk * 32767).astype(np.int16),
|
||||
)
|
||||
except:
|
||||
print(inp_path,"->fail->",traceback.format_exc())
|
||||
return "执行完毕，请检查输出文件"
|
||||
|
||||
print(slice(*sys.argv[1:]))
|
||||
|
||||
@ -1,261 +1,261 @@
|
||||
import numpy as np
|
||||
|
||||
|
||||
# This function is obtained from librosa.
|
||||
def get_rms(
|
||||
y,
|
||||
frame_length=2048,
|
||||
hop_length=512,
|
||||
pad_mode="constant",
|
||||
):
|
||||
padding = (int(frame_length // 2), int(frame_length // 2))
|
||||
y = np.pad(y, padding, mode=pad_mode)
|
||||
|
||||
axis = -1
|
||||
# put our new within-frame axis at the end for now
|
||||
out_strides = y.strides + tuple([y.strides[axis]])
|
||||
# Reduce the shape on the framing axis
|
||||
x_shape_trimmed = list(y.shape)
|
||||
x_shape_trimmed[axis] -= frame_length - 1
|
||||
out_shape = tuple(x_shape_trimmed) + tuple([frame_length])
|
||||
xw = np.lib.stride_tricks.as_strided(y, shape=out_shape, strides=out_strides)
|
||||
if axis < 0:
|
||||
target_axis = axis - 1
|
||||
else:
|
||||
target_axis = axis + 1
|
||||
xw = np.moveaxis(xw, -1, target_axis)
|
||||
# Downsample along the target axis
|
||||
slices = [slice(None)] * xw.ndim
|
||||
slices[axis] = slice(0, None, hop_length)
|
||||
x = xw[tuple(slices)]
|
||||
|
||||
# Calculate power
|
||||
power = np.mean(np.abs(x) ** 2, axis=-2, keepdims=True)
|
||||
|
||||
return np.sqrt(power)
|
||||
|
||||
|
||||
class Slicer:
|
||||
def __init__(
|
||||
self,
|
||||
sr: int,
|
||||
threshold: float = -40.0,
|
||||
min_length: int = 5000,
|
||||
min_interval: int = 300,
|
||||
hop_size: int = 20,
|
||||
max_sil_kept: int = 5000,
|
||||
):
|
||||
if not min_length >= min_interval >= hop_size:
|
||||
raise ValueError(
|
||||
"The following condition must be satisfied: min_length >= min_interval >= hop_size"
|
||||
)
|
||||
if not max_sil_kept >= hop_size:
|
||||
raise ValueError(
|
||||
"The following condition must be satisfied: max_sil_kept >= hop_size"
|
||||
)
|
||||
min_interval = sr * min_interval / 1000
|
||||
self.threshold = 10 ** (threshold / 20.0)
|
||||
self.hop_size = round(sr * hop_size / 1000)
|
||||
self.win_size = min(round(min_interval), 4 * self.hop_size)
|
||||
self.min_length = round(sr * min_length / 1000 / self.hop_size)
|
||||
self.min_interval = round(min_interval / self.hop_size)
|
||||
self.max_sil_kept = round(sr * max_sil_kept / 1000 / self.hop_size)
|
||||
|
||||
def _apply_slice(self, waveform, begin, end):
|
||||
if len(waveform.shape) > 1:
|
||||
return waveform[
|
||||
:, begin * self.hop_size : min(waveform.shape[1], end * self.hop_size)
|
||||
]
|
||||
else:
|
||||
return waveform[
|
||||
begin * self.hop_size : min(waveform.shape[0], end * self.hop_size)
|
||||
]
|
||||
|
||||
# @timeit
|
||||
def slice(self, waveform):
|
||||
if len(waveform.shape) > 1:
|
||||
samples = waveform.mean(axis=0)
|
||||
else:
|
||||
samples = waveform
|
||||
if samples.shape[0] <= self.min_length:
|
||||
return [waveform]
|
||||
rms_list = get_rms(
|
||||
y=samples, frame_length=self.win_size, hop_length=self.hop_size
|
||||
).squeeze(0)
|
||||
sil_tags = []
|
||||
silence_start = None
|
||||
clip_start = 0
|
||||
for i, rms in enumerate(rms_list):
|
||||
# Keep looping while frame is silent.
|
||||
if rms < self.threshold:
|
||||
# Record start of silent frames.
|
||||
if silence_start is None:
|
||||
silence_start = i
|
||||
continue
|
||||
# Keep looping while frame is not silent and silence start has not been recorded.
|
||||
if silence_start is None:
|
||||
continue
|
||||
# Clear recorded silence start if interval is not enough or clip is too short
|
||||
is_leading_silence = silence_start == 0 and i > self.max_sil_kept
|
||||
need_slice_middle = (
|
||||
i - silence_start >= self.min_interval
|
||||
and i - clip_start >= self.min_length
|
||||
)
|
||||
if not is_leading_silence and not need_slice_middle:
|
||||
silence_start = None
|
||||
continue
|
||||
# Need slicing. Record the range of silent frames to be removed.
|
||||
if i - silence_start <= self.max_sil_kept:
|
||||
pos = rms_list[silence_start : i + 1].argmin() + silence_start
|
||||
if silence_start == 0:
|
||||
sil_tags.append((0, pos))
|
||||
else:
|
||||
sil_tags.append((pos, pos))
|
||||
clip_start = pos
|
||||
elif i - silence_start <= self.max_sil_kept * 2:
|
||||
pos = rms_list[
|
||||
i - self.max_sil_kept : silence_start + self.max_sil_kept + 1
|
||||
].argmin()
|
||||
pos += i - self.max_sil_kept
|
||||
pos_l = (
|
||||
rms_list[
|
||||
silence_start : silence_start + self.max_sil_kept + 1
|
||||
].argmin()
|
||||
+ silence_start
|
||||
)
|
||||
pos_r = (
|
||||
rms_list[i - self.max_sil_kept : i + 1].argmin()
|
||||
+ i
|
||||
- self.max_sil_kept
|
||||
)
|
||||
if silence_start == 0:
|
||||
sil_tags.append((0, pos_r))
|
||||
clip_start = pos_r
|
||||
else:
|
||||
sil_tags.append((min(pos_l, pos), max(pos_r, pos)))
|
||||
clip_start = max(pos_r, pos)
|
||||
else:
|
||||
pos_l = (
|
||||
rms_list[
|
||||
silence_start : silence_start + self.max_sil_kept + 1
|
||||
].argmin()
|
||||
+ silence_start
|
||||
)
|
||||
pos_r = (
|
||||
rms_list[i - self.max_sil_kept : i + 1].argmin()
|
||||
+ i
|
||||
- self.max_sil_kept
|
||||
)
|
||||
if silence_start == 0:
|
||||
sil_tags.append((0, pos_r))
|
||||
else:
|
||||
sil_tags.append((pos_l, pos_r))
|
||||
clip_start = pos_r
|
||||
silence_start = None
|
||||
# Deal with trailing silence.
|
||||
total_frames = rms_list.shape[0]
|
||||
if (
|
||||
silence_start is not None
|
||||
and total_frames - silence_start >= self.min_interval
|
||||
):
|
||||
silence_end = min(total_frames, silence_start + self.max_sil_kept)
|
||||
pos = rms_list[silence_start : silence_end + 1].argmin() + silence_start
|
||||
sil_tags.append((pos, total_frames + 1))
|
||||
# Apply and return slices.
|
||||
####音频+起始时间+终止时间
|
||||
if len(sil_tags) == 0:
|
||||
return [[waveform,0,int(total_frames*self.hop_size)]]
|
||||
else:
|
||||
chunks = []
|
||||
if sil_tags[0][0] > 0:
|
||||
chunks.append([self._apply_slice(waveform, 0, sil_tags[0][0]),0,int(sil_tags[0][0]*self.hop_size)])
|
||||
for i in range(len(sil_tags) - 1):
|
||||
chunks.append(
|
||||
[self._apply_slice(waveform, sil_tags[i][1], sil_tags[i + 1][0]),int(sil_tags[i][1]*self.hop_size),int(sil_tags[i + 1][0]*self.hop_size)]
|
||||
)
|
||||
if sil_tags[-1][1] < total_frames:
|
||||
chunks.append(
|
||||
[self._apply_slice(waveform, sil_tags[-1][1], total_frames),int(sil_tags[-1][1]*self.hop_size),int(total_frames*self.hop_size)]
|
||||
)
|
||||
return chunks
|
||||
|
||||
|
||||
def main():
|
||||
import os.path
|
||||
from argparse import ArgumentParser
|
||||
|
||||
import librosa
|
||||
import soundfile
|
||||
|
||||
parser = ArgumentParser()
|
||||
parser.add_argument("audio", type=str, help="The audio to be sliced")
|
||||
parser.add_argument(
|
||||
"--out", type=str, help="Output directory of the sliced audio clips"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--db_thresh",
|
||||
type=float,
|
||||
required=False,
|
||||
default=-40,
|
||||
help="The dB threshold for silence detection",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--min_length",
|
||||
type=int,
|
||||
required=False,
|
||||
default=5000,
|
||||
help="The minimum milliseconds required for each sliced audio clip",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--min_interval",
|
||||
type=int,
|
||||
required=False,
|
||||
default=300,
|
||||
help="The minimum milliseconds for a silence part to be sliced",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--hop_size",
|
||||
type=int,
|
||||
required=False,
|
||||
default=10,
|
||||
help="Frame length in milliseconds",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--max_sil_kept",
|
||||
type=int,
|
||||
required=False,
|
||||
default=500,
|
||||
help="The maximum silence length kept around the sliced clip, presented in milliseconds",
|
||||
)
|
||||
args = parser.parse_args()
|
||||
out = args.out
|
||||
if out is None:
|
||||
out = os.path.dirname(os.path.abspath(args.audio))
|
||||
audio, sr = librosa.load(args.audio, sr=None, mono=False)
|
||||
slicer = Slicer(
|
||||
sr=sr,
|
||||
threshold=args.db_thresh,
|
||||
min_length=args.min_length,
|
||||
min_interval=args.min_interval,
|
||||
hop_size=args.hop_size,
|
||||
max_sil_kept=args.max_sil_kept,
|
||||
)
|
||||
chunks = slicer.slice(audio)
|
||||
if not os.path.exists(out):
|
||||
os.makedirs(out)
|
||||
for i, chunk in enumerate(chunks):
|
||||
if len(chunk.shape) > 1:
|
||||
chunk = chunk.T
|
||||
soundfile.write(
|
||||
os.path.join(
|
||||
out,
|
||||
f"%s_%d.wav"
|
||||
% (os.path.basename(args.audio).rsplit(".", maxsplit=1)[0], i),
|
||||
),
|
||||
chunk,
|
||||
sr,
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
import numpy as np
|
||||
|
||||
|
||||
# This function is obtained from librosa.
|
||||
def get_rms(
|
||||
y,
|
||||
frame_length=2048,
|
||||
hop_length=512,
|
||||
pad_mode="constant",
|
||||
):
|
||||
padding = (int(frame_length // 2), int(frame_length // 2))
|
||||
y = np.pad(y, padding, mode=pad_mode)
|
||||
|
||||
axis = -1
|
||||
# put our new within-frame axis at the end for now
|
||||
out_strides = y.strides + tuple([y.strides[axis]])
|
||||
# Reduce the shape on the framing axis
|
||||
x_shape_trimmed = list(y.shape)
|
||||
x_shape_trimmed[axis] -= frame_length - 1
|
||||
out_shape = tuple(x_shape_trimmed) + tuple([frame_length])
|
||||
xw = np.lib.stride_tricks.as_strided(y, shape=out_shape, strides=out_strides)
|
||||
if axis < 0:
|
||||
target_axis = axis - 1
|
||||
else:
|
||||
target_axis = axis + 1
|
||||
xw = np.moveaxis(xw, -1, target_axis)
|
||||
# Downsample along the target axis
|
||||
slices = [slice(None)] * xw.ndim
|
||||
slices[axis] = slice(0, None, hop_length)
|
||||
x = xw[tuple(slices)]
|
||||
|
||||
# Calculate power
|
||||
power = np.mean(np.abs(x) ** 2, axis=-2, keepdims=True)
|
||||
|
||||
return np.sqrt(power)
|
||||
|
||||
|
||||
class Slicer:
|
||||
def __init__(
|
||||
self,
|
||||
sr: int,
|
||||
threshold: float = -40.0,
|
||||
min_length: int = 5000,
|
||||
min_interval: int = 300,
|
||||
hop_size: int = 20,
|
||||
max_sil_kept: int = 5000,
|
||||
):
|
||||
if not min_length >= min_interval >= hop_size:
|
||||
raise ValueError(
|
||||
"The following condition must be satisfied: min_length >= min_interval >= hop_size"
|
||||
)
|
||||
if not max_sil_kept >= hop_size:
|
||||
raise ValueError(
|
||||
"The following condition must be satisfied: max_sil_kept >= hop_size"
|
||||
)
|
||||
min_interval = sr * min_interval / 1000
|
||||
self.threshold = 10 ** (threshold / 20.0)
|
||||
self.hop_size = round(sr * hop_size / 1000)
|
||||
self.win_size = min(round(min_interval), 4 * self.hop_size)
|
||||
self.min_length = round(sr * min_length / 1000 / self.hop_size)
|
||||
self.min_interval = round(min_interval / self.hop_size)
|
||||
self.max_sil_kept = round(sr * max_sil_kept / 1000 / self.hop_size)
|
||||
|
||||
def _apply_slice(self, waveform, begin, end):
|
||||
if len(waveform.shape) > 1:
|
||||
return waveform[
|
||||
:, begin * self.hop_size : min(waveform.shape[1], end * self.hop_size)
|
||||
]
|
||||
else:
|
||||
return waveform[
|
||||
begin * self.hop_size : min(waveform.shape[0], end * self.hop_size)
|
||||
]
|
||||
|
||||
# @timeit
|
||||
def slice(self, waveform):
|
||||
if len(waveform.shape) > 1:
|
||||
samples = waveform.mean(axis=0)
|
||||
else:
|
||||
samples = waveform
|
||||
if samples.shape[0] <= self.min_length:
|
||||
return [waveform]
|
||||
rms_list = get_rms(
|
||||
y=samples, frame_length=self.win_size, hop_length=self.hop_size
|
||||
).squeeze(0)
|
||||
sil_tags = []
|
||||
silence_start = None
|
||||
clip_start = 0
|
||||
for i, rms in enumerate(rms_list):
|
||||
# Keep looping while frame is silent.
|
||||
if rms < self.threshold:
|
||||
# Record start of silent frames.
|
||||
if silence_start is None:
|
||||
silence_start = i
|
||||
continue
|
||||
# Keep looping while frame is not silent and silence start has not been recorded.
|
||||
if silence_start is None:
|
||||
continue
|
||||
# Clear recorded silence start if interval is not enough or clip is too short
|
||||
is_leading_silence = silence_start == 0 and i > self.max_sil_kept
|
||||
need_slice_middle = (
|
||||
i - silence_start >= self.min_interval
|
||||
and i - clip_start >= self.min_length
|
||||
)
|
||||
if not is_leading_silence and not need_slice_middle:
|
||||
silence_start = None
|
||||
continue
|
||||
# Need slicing. Record the range of silent frames to be removed.
|
||||
if i - silence_start <= self.max_sil_kept:
|
||||
pos = rms_list[silence_start : i + 1].argmin() + silence_start
|
||||
if silence_start == 0:
|
||||
sil_tags.append((0, pos))
|
||||
else:
|
||||
sil_tags.append((pos, pos))
|
||||
clip_start = pos
|
||||
elif i - silence_start <= self.max_sil_kept * 2:
|
||||
pos = rms_list[
|
||||
i - self.max_sil_kept : silence_start + self.max_sil_kept + 1
|
||||
].argmin()
|
||||
pos += i - self.max_sil_kept
|
||||
pos_l = (
|
||||
rms_list[
|
||||
silence_start : silence_start + self.max_sil_kept + 1
|
||||
].argmin()
|
||||
+ silence_start
|
||||
)
|
||||
pos_r = (
|
||||
rms_list[i - self.max_sil_kept : i + 1].argmin()
|
||||
+ i
|
||||
- self.max_sil_kept
|
||||
)
|
||||
if silence_start == 0:
|
||||
sil_tags.append((0, pos_r))
|
||||
clip_start = pos_r
|
||||
else:
|
||||
sil_tags.append((min(pos_l, pos), max(pos_r, pos)))
|
||||
clip_start = max(pos_r, pos)
|
||||
else:
|
||||
pos_l = (
|
||||
rms_list[
|
||||
silence_start : silence_start + self.max_sil_kept + 1
|
||||
].argmin()
|
||||
+ silence_start
|
||||
)
|
||||
pos_r = (
|
||||
rms_list[i - self.max_sil_kept : i + 1].argmin()
|
||||
+ i
|
||||
- self.max_sil_kept
|
||||
)
|
||||
if silence_start == 0:
|
||||
sil_tags.append((0, pos_r))
|
||||
else:
|
||||
sil_tags.append((pos_l, pos_r))
|
||||
clip_start = pos_r
|
||||
silence_start = None
|
||||
# Deal with trailing silence.
|
||||
total_frames = rms_list.shape[0]
|
||||
if (
|
||||
silence_start is not None
|
||||
and total_frames - silence_start >= self.min_interval
|
||||
):
|
||||
silence_end = min(total_frames, silence_start + self.max_sil_kept)
|
||||
pos = rms_list[silence_start : silence_end + 1].argmin() + silence_start
|
||||
sil_tags.append((pos, total_frames + 1))
|
||||
# Apply and return slices.
|
||||
####音频+起始时间+终止时间
|
||||
if len(sil_tags) == 0:
|
||||
return [[waveform,0,int(total_frames*self.hop_size)]]
|
||||
else:
|
||||
chunks = []
|
||||
if sil_tags[0][0] > 0:
|
||||
chunks.append([self._apply_slice(waveform, 0, sil_tags[0][0]),0,int(sil_tags[0][0]*self.hop_size)])
|
||||
for i in range(len(sil_tags) - 1):
|
||||
chunks.append(
|
||||
[self._apply_slice(waveform, sil_tags[i][1], sil_tags[i + 1][0]),int(sil_tags[i][1]*self.hop_size),int(sil_tags[i + 1][0]*self.hop_size)]
|
||||
)
|
||||
if sil_tags[-1][1] < total_frames:
|
||||
chunks.append(
|
||||
[self._apply_slice(waveform, sil_tags[-1][1], total_frames),int(sil_tags[-1][1]*self.hop_size),int(total_frames*self.hop_size)]
|
||||
)
|
||||
return chunks
|
||||
|
||||
|
||||
def main():
|
||||
import os.path
|
||||
from argparse import ArgumentParser
|
||||
|
||||
import librosa
|
||||
import soundfile
|
||||
|
||||
parser = ArgumentParser()
|
||||
parser.add_argument("audio", type=str, help="The audio to be sliced")
|
||||
parser.add_argument(
|
||||
"--out", type=str, help="Output directory of the sliced audio clips"
|
||||
)
|
||||
parser.add_argument(
|
||||
"--db_thresh",
|
||||
type=float,
|
||||
required=False,
|
||||
default=-40,
|
||||
help="The dB threshold for silence detection",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--min_length",
|
||||
type=int,
|
||||
required=False,
|
||||
default=5000,
|
||||
help="The minimum milliseconds required for each sliced audio clip",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--min_interval",
|
||||
type=int,
|
||||
required=False,
|
||||
default=300,
|
||||
help="The minimum milliseconds for a silence part to be sliced",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--hop_size",
|
||||
type=int,
|
||||
required=False,
|
||||
default=10,
|
||||
help="Frame length in milliseconds",
|
||||
)
|
||||
parser.add_argument(
|
||||
"--max_sil_kept",
|
||||
type=int,
|
||||
required=False,
|
||||
default=500,
|
||||
help="The maximum silence length kept around the sliced clip, presented in milliseconds",
|
||||
)
|
||||
args = parser.parse_args()
|
||||
out = args.out
|
||||
if out is None:
|
||||
out = os.path.dirname(os.path.abspath(args.audio))
|
||||
audio, sr = librosa.load(args.audio, sr=None, mono=False)
|
||||
slicer = Slicer(
|
||||
sr=sr,
|
||||
threshold=args.db_thresh,
|
||||
min_length=args.min_length,
|
||||
min_interval=args.min_interval,
|
||||
hop_size=args.hop_size,
|
||||
max_sil_kept=args.max_sil_kept,
|
||||
)
|
||||
chunks = slicer.slice(audio)
|
||||
if not os.path.exists(out):
|
||||
os.makedirs(out)
|
||||
for i, chunk in enumerate(chunks):
|
||||
if len(chunk.shape) > 1:
|
||||
chunk = chunk.T
|
||||
soundfile.write(
|
||||
os.path.join(
|
||||
out,
|
||||
f"%s_%d.wav"
|
||||
% (os.path.basename(args.audio).rsplit(".", maxsplit=1)[0], i),
|
||||
),
|
||||
chunk,
|
||||
sr,
|
||||
)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
@ -1,54 +1,54 @@
|
||||
{
|
||||
"bins": 672,
|
||||
"unstable_bins": 8,
|
||||
"reduction_bins": 530,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 640,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 85,
|
||||
"lpf_start": 25,
|
||||
"lpf_stop": 53,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 320,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 87,
|
||||
"hpf_start": 25,
|
||||
"hpf_stop": 12,
|
||||
"lpf_start": 31,
|
||||
"lpf_stop": 62,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 14700,
|
||||
"hl": 160,
|
||||
"n_fft": 512,
|
||||
"crop_start": 17,
|
||||
"crop_stop": 216,
|
||||
"hpf_start": 48,
|
||||
"hpf_stop": 24,
|
||||
"lpf_start": 139,
|
||||
"lpf_stop": 210,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 480,
|
||||
"n_fft": 960,
|
||||
"crop_start": 78,
|
||||
"crop_stop": 383,
|
||||
"hpf_start": 130,
|
||||
"hpf_stop": 86,
|
||||
"res_type": "kaiser_fast"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 668,
|
||||
"pre_filter_stop": 672
|
||||
{
|
||||
"bins": 672,
|
||||
"unstable_bins": 8,
|
||||
"reduction_bins": 530,
|
||||
"band": {
|
||||
"1": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 640,
|
||||
"crop_start": 0,
|
||||
"crop_stop": 85,
|
||||
"lpf_start": 25,
|
||||
"lpf_stop": 53,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"2": {
|
||||
"sr": 7350,
|
||||
"hl": 80,
|
||||
"n_fft": 320,
|
||||
"crop_start": 4,
|
||||
"crop_stop": 87,
|
||||
"hpf_start": 25,
|
||||
"hpf_stop": 12,
|
||||
"lpf_start": 31,
|
||||
"lpf_stop": 62,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"3": {
|
||||
"sr": 14700,
|
||||
"hl": 160,
|
||||
"n_fft": 512,
|
||||
"crop_start": 17,
|
||||
"crop_stop": 216,
|
||||
"hpf_start": 48,
|
||||
"hpf_stop": 24,
|
||||
"lpf_start": 139,
|
||||
"lpf_stop": 210,
|
||||
"res_type": "polyphase"
|
||||
},
|
||||
"4": {
|
||||
"sr": 44100,
|
||||
"hl": 480,
|
||||
"n_fft": 960,
|
||||
"crop_start": 78,
|
||||
"crop_stop": 383,
|
||||
"hpf_start": 130,
|
||||
"hpf_stop": 86,
|
||||
"res_type": "kaiser_fast"
|
||||
}
|
||||
},
|
||||
"sr": 44100,
|
||||
"pre_filter_start": 668,
|
||||
"pre_filter_stop": 672
|
||||
}
|
||||
1
MANIFEST.in
Normal file
1
MANIFEST.in
Normal file
@ -0,0 +1 @@
|
||||
include GPT_SoVITS/configs/*
|
||||
@ -1,3 +1,6 @@
|
||||
# Jarod's NOTE
|
||||
Working on turning this into a package. Right now, the API *does in fact* work to make requests to and this can be installed. Will get instrucitons out sometime in the future.
|
||||
|
||||
<div align="center">
|
||||
|
||||
|
||||
|
||||
63
pyproject.toml
Normal file
63
pyproject.toml
Normal file
@ -0,0 +1,63 @@
|
||||
[build-system]
|
||||
requires = ["setuptools >= 61.0", "setuptools-scm>=8.0"]
|
||||
build-backend = "setuptools.build_meta"
|
||||
|
||||
[project]
|
||||
name = "GPT_SoVITS"
|
||||
dynamic = ["version"]
|
||||
description = ""
|
||||
readme = "README.md"
|
||||
license = {text = "MIT License"}
|
||||
classifiers = [
|
||||
"License :: OSI Approved :: MIT License",
|
||||
"Operating System :: OS Independent",
|
||||
"Programming Language :: Python :: 3",
|
||||
]
|
||||
dependencies = [
|
||||
"numpy==1.26.4",
|
||||
"scipy",
|
||||
"tensorboard",
|
||||
"librosa==0.9.2",
|
||||
"numba",
|
||||
"pytorch-lightning",
|
||||
"gradio>=4.0,<=4.24.0",
|
||||
"ffmpeg-python",
|
||||
"onnxruntime; sys_platform == 'darwin'",
|
||||
"onnxruntime-gpu; sys_platform != 'darwin'",
|
||||
"tqdm",
|
||||
"funasr==1.0.27",
|
||||
"cn2an",
|
||||
"pypinyin",
|
||||
"pyopenjtalk>=0.3.4",
|
||||
"g2p_en",
|
||||
"torchaudio",
|
||||
"modelscope==1.10.0",
|
||||
"sentencepiece",
|
||||
"transformers",
|
||||
"chardet",
|
||||
"PyYAML",
|
||||
"psutil",
|
||||
"jieba_fast",
|
||||
"jieba",
|
||||
"LangSegment>=0.2.0",
|
||||
"Faster_Whisper",
|
||||
"wordsegment",
|
||||
"rotary_embedding_torch",
|
||||
"pyjyutping",
|
||||
"g2pk2",
|
||||
"ko_pron",
|
||||
"opencc; sys_platform != 'linux'",
|
||||
"opencc==1.1.1; sys_platform == 'linux'",
|
||||
"python_mecab_ko; sys_platform != 'win32'",
|
||||
"fastapi<0.112.2",
|
||||
]
|
||||
|
||||
|
||||
[project.urls]
|
||||
Homepage = "https://github.com/RVC-Boss/GPT-SoVITS"
|
||||
|
||||
[project.scripts]
|
||||
"gpt_sovits_api" = "GPT_SoVITS.api_v2:main"
|
||||
|
||||
[tool.setuptools.packages.find]
|
||||
include = ["GPT_SoVITS"]
|
||||
@ -1,8 +1,8 @@
|
||||
numpy==1.23.4
|
||||
numpy==1.26.4
|
||||
scipy
|
||||
tensorboard
|
||||
librosa==0.9.2
|
||||
numba==0.56.4
|
||||
numba
|
||||
pytorch-lightning
|
||||
gradio>=4.0,<=4.24.0
|
||||
ffmpeg-python
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user