Kit-Lemonfoot commited on
Commit
f646510
1 Parent(s): 4339a8b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +16 -1
README.md CHANGED
@@ -22,5 +22,20 @@ This includes the following:
22
  - Every dataset used in [my Style BertVITS2 models](https://huggingface.co/spaces/Kit-Lemonfoot/Hololive-Style-Bert-VITS2)
23
  - Every dataset used in [my GPT-SoVITS models](https://huggingface.co/Kit-Lemonfoot/kitlemonfoot_gptsovits_models)
24
  - Multiple datasets submitted to 15.ai (RIP)
25
-
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
26
  All data was taken from multiple publicly available YouTube sources. **These datasets are not to be used for commercial purposes.**
 
22
  - Every dataset used in [my Style BertVITS2 models](https://huggingface.co/spaces/Kit-Lemonfoot/Hololive-Style-Bert-VITS2)
23
  - Every dataset used in [my GPT-SoVITS models](https://huggingface.co/Kit-Lemonfoot/kitlemonfoot_gptsovits_models)
24
  - Multiple datasets submitted to 15.ai (RIP)
25
+
26
+ ## Reference
27
+ Hand-transcribed datasets were built for 15.ai and as such are in the LJSpeech transcription format.
28
+ ```
29
+ SPEAKERNAME/filename1.wav|transcription
30
+ SPEAKERNAME/filename2.wav|transcription
31
+ SPEAKERNAME/filename3.wav|transcription
32
+ ```
33
+ Whisper-transcribed datasets are in the VITS transcription format as seen in BertVITS2 and its' contemporaries.
34
+ ```
35
+ filename1.wav|SPEAKERNAME|LANGUAGE|transcription
36
+ filename2.wav|SPEAKERNAME|LANGUAGE|transcription
37
+ filename3.wav|SPEAKERNAME|LANGUAGE|transcription
38
+ ```
39
+ While some non-transcribed datasets may not be clipped, they have had their silences removed. Please be mindful of this while clipping.
40
+ ## Other Information
41
  All data was taken from multiple publicly available YouTube sources. **These datasets are not to be used for commercial purposes.**