Update README.md
Browse files
README.md
CHANGED
|
@@ -2,4 +2,21 @@
|
|
| 2 |
language:
|
| 3 |
- fr
|
| 4 |
pipeline_tag: text-to-speech
|
| 5 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
language:
|
| 3 |
- fr
|
| 4 |
pipeline_tag: text-to-speech
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
This model was trained using part of SIWIS dataset and CSS10 dataset, only in french with two voices (F and M),
|
| 8 |
+
It was trained for finetuning task in language similar than french.
|
| 9 |
+
|
| 10 |
+
SIWIS dataset: https://datashare.ed.ac.uk/handle/10283/2353
|
| 11 |
+
CSS10 dataset : https://github.com/Kyubyong/css10
|
| 12 |
+
|
| 13 |
+
The first 3 parts of the SIWIS dataset, corresponding to the previously segmented female voice,
|
| 14 |
+
have been taken from the first dataset. The second dataset has been fully used and corresponds to the male voice.
|
| 15 |
+
In both cases the sampling frequency (22050) and the wav format encoding (pmc 16) have been adjusted.
|
| 16 |
+
In addition, the transcriptions have been adapted by adjusting the format with the ljspeech dataset.
|
| 17 |
+
|
| 18 |
+
The alphabet used is:
|
| 19 |
+
|
| 20 |
+
!"&'()+,-./0123456789:;?ABCDEFGHIJKLMNOPQRSTUVWXYZ_abcdefghijklmnopqrstuvwxyz| «°»ÀÂÇÈÉÊÎÏÔÙÛàâæçèéêëîïôöùúûüœ–’“”…
|
| 21 |
+
|
| 22 |
+
|