To Latest

README.md

@@ -8,6 +8,8 @@
 [![Open In Colab](https://img.shields.io/badge/Colab-F9AB00?style=for-the-badge&logo=googlecolab&color=525252)](https://colab.research.google.com/github/svc-develop-team/so-vits-svc/blob/4.1-Stable/sovits4_for_colab.ipynb)
 [![Licence](https://img.shields.io/badge/LICENSE-AGPL3.0-green.svg?style=for-the-badge)](https://github.com/svc-develop-team/so-vits-svc/blob/4.1-Stable/LICENSE)
 
+This round of limited time update is coming to an end, the warehouse will enter the Archieve state, please know
+
 </div>
 
 > ✨ A studio that contains visible f0 editor, speaker mix timeline editor and other features (Where the Onnx models are used) : [MoeVoiceStudio](https://github.com/NaruseMioShirakana/MoeVoiceStudio)
@@ -318,7 +320,7 @@ fcpe
 
 If the training set is too noisy,it is recommended to use `crepe` to handle f0
 
-If the f0_predictor parameter is omitted, the default value is `dio`
+If the f0_predictor parameter is omitted, the default value is `rmvpe`
 
 If you want shallow diffusion (optional), you need to add the `--use_diff` parameter, for example:
 
@@ -331,7 +333,7 @@ python preprocess_hubert_f0.py --f0_predictor dio --use_diff
 If your dataset is pretty large,you can increase the param `--num_processes` like that:
 
 ```shell
-python preprocess_hubert_f0.py --speech_encoder vec768l12 --vol_aug --num_processes 8
+python preprocess_hubert_f0.py --f0_predictor dio --num_processes 8
 ```
 All the worker will be assigned to different GPU if you have more than one GPUs.
 

README_zh_CN.md

@@ -8,6 +8,8 @@
 [![在Google Cloab中打开](https://img.shields.io/badge/Colab-F9AB00?style=for-the-badge&logo=googlecolab&color=525252)](https://colab.research.google.com/github/svc-develop-team/so-vits-svc/blob/4.1-Stable/sovits4_for_colab.ipynb)
 [![LICENSE](https://img.shields.io/badge/LICENSE-AGPL3.0-green.svg?style=for-the-badge)](https://github.com/svc-develop-team/so-vits-svc/blob/4.1-Stable/LICENSE)
 
+本轮限时更新即将结束，仓库将进入Archieve状态，望周知
+
 </div>
 
 
@@ -323,7 +325,7 @@ fcpe
 
 如果训练集过于嘈杂，请使用 crepe 处理 f0
 
-如果省略 f0_predictor 参数，默认值为 dio
+如果省略 f0_predictor 参数，默认值为 rmvpe
 
 尚若需要浅扩散功能（可选），需要增加--use_diff 参数，比如
 

onnx_export.py

@@ -1,8 +1,11 @@
+import argparse
 import json
+
 import torch
+
 import utils
 from onnxexport.model_onnx_speaker_mix import SynthesizerTrn
-import argparse
+
 parser = argparse.ArgumentParser(description='SoVitsSvc OnnxExport')
 
 def OnnxExport(path=None):

onnxexport/model_onnx_speaker_mix.py

@@ -1,14 +1,13 @@
 import torch
 from torch import nn
 from torch.nn import Conv1d, Conv2d
 from torch.nn import functional as F
 from torch.nn.utils import spectral_norm, weight_norm
 
 import modules.attentions as attentions
 import modules.commons as commons
 import modules.modules as modules
 import utils
-from modules.commons import get_padding
 from utils import f0_to_coarse
 
 

preprocess_hubert_f0.py

@@ -137,7 +137,7 @@ def parallel_process(filenames, num_processes, f0p, diff, mel_extractor, device)
         '--use_diff',action='store_true', help='Whether to use the diffusion model'
     )
     parser.add_argument(
-        '--f0_predictor', type=str, default="dio", help='Select F0 predictor, can select crepe,pm,dio,harvest,rmvpe,fcpe|default: pm(note: crepe is original F0 using mean filter)'
+        '--f0_predictor', type=str, default="rmvpe", help='Select F0 predictor, can select crepe,pm,dio,harvest,rmvpe,fcpe|default: pm(note: crepe is original F0 using mean filter)'
     )
     parser.add_argument(
         '--num_processes', type=int, default=1, help='You are advised to set the number of processes to the same as the number of CPU cores'