Voice data <= 10 mins can also be used to train a good VC model!
audio-analysischangeconversational-aiconversionconverterretrieval-modelretrieve-datarvcsovitsso-vits-svcvcvitsvoicevoice-conversionvoiceconversionvoice-converter
configs | ||
infer | ||
infer_pack | ||
pretrained | ||
train | ||
uvr5_pack | ||
uvr5_weights | ||
weights | ||
.gitignore | ||
config.py | ||
extract_f0_print.py | ||
extract_feature_print.py | ||
infer_uvr5.py | ||
infer-web.py | ||
LICENSE | ||
my_utils.py | ||
README.md | ||
requirements.txt | ||
slicer2.py | ||
train_nsf_sim_cache_sid_load_pretrain.py | ||
trainset_preprocess_pipeline_print.py | ||
vc_infer_pipeline.py | ||
使用需遵守的协议-LICENSE.txt | ||
小白简易教程.doc |
Retrieval-based-Voice-Conversion-WebUI
缺失的2个文件夹和2个文件:
hubert_base.pt
ffmpeg(自己确保ffmpeg命令能执行就行)
pretrained文件夹
uvr5_weights文件夹
文件太大github传不动,去huggingface上下https://huggingface.co/lj1995/VoiceConversionWebUI/tree/main
当然你也可以直接看看RVC-beta.7z这个文件→_→
按照requirements.txt用pip装好环境,python infer-web.py就能用了
根据经验,librosa numpy和numba三个包最好写死版本否则容易有坑,其他的包版本不太重要
宣传视频:https://www.bilibili.com/video/BV1pm4y1z7Gm/
教程见小白简易教程.doc