I have surveyed a few of the languages and one thing I have not found is a good description of how the dev/train/test sets are split. I would be interested to know what the algorithm is. For instance, it looks like early on, where there isn’t a large amount of data in a language, the test and dev se…

How are the dev/test/train datasets split?

kdavis (kdavis) March 7, 2019, 8:10pm 4

They will be re-generated using the CorporaCreator which will not maintain the previous splits.

1 Like

Topic		Replies	Views
Common Voice datasets (Mandarin zh-tw) Common Voice dataset	2	918	May 23, 2019
Dataset split best practices? Common Voice feedback , dataset	23	4768	December 23, 2019
Speaker ID split between train/test/dev Common Voice dataset	4	1023	February 15, 2019
Common Voice v1 corpus design problems, overlapping train/test/dev sentences Common Voice dataset	2	2222	April 3, 2018
Are splits currently still regenerated per version? Common Voice dataset	6	778	June 27, 2023