Update README_zh_CN.md
This commit is contained in:
parent
968c508836
commit
2695419671
|
@ -170,6 +170,8 @@ dataset_raw
|
|||
|
||||
切完之后手动删除过长过短的音频
|
||||
|
||||
**如果你使用Whisper-ppg声音编码器进行训练,所有的切片长度必须小于30s**
|
||||
|
||||
### 1. 重采样至44100Hz单声道
|
||||
|
||||
```shell
|
||||
|
@ -182,7 +184,7 @@ python resample.py
|
|||
python preprocess_flist_config.py --speech_encoder whisper-ppg
|
||||
```
|
||||
|
||||
speech_encoder拥有三个选择
|
||||
speech_encoder拥有四个选择
|
||||
|
||||
```
|
||||
vec768l12
|
||||
|
|
Loading…
Reference in New Issue