How to use multiple csv files in train directory

[JRMeyer]

>>> elpimous_robot
[December 3, 2017, 10:52pm]

Trying the deepspeech voice-corpus-tools, I decided to augment my
TRAIN dir, with tempo and speed (±5%).

How could I tell to deepspeech to use more than one csv file ? slash
I have 3 CSV files, with same wav names !!! slash
Thanks.

[This is an archived TTS discussion thread from discourse.mozilla.org/t/how-to-use-multiple-csv-files-in-train-directory]

[JRMeyer]

>>> elpimous_robot
[December 10, 2017, 12:13pm]

Thanks Reuben.

--train_files /first.csv, /next.csv *slash*

just use some csv files, separed by a comma !

note : here, both csv files have same wav name (objective : duplicate
my train voices with voice-corpus-tools, and improve accuracy with
some minor changes in sounds)

works perfectly.

[Archived Post]