How to do the training

Deepspeech training, what does it depend on?
of the number of audios, and the amount of text in the text file ?,
or with the same files and texts put him to train many times