Why mp3 files? and why tsv? and how to test my trained model?

hi, first question :
Common voice data has tsv and mp3 file format. what if I have a CSV file instead of tsv. and my audio files are wave format instead of mp3.
Does bin/import_cv2.py command accept my data?

Second question:
Now I have trained my model with common voice data in persian. how can I test my model? feed it with my own voice?
I couldn’t find this command in flags page and doc.

Have you read documentation and other importers? It does not looks like you did, because the basics requirements for deepspeech are wave + csv.


1 Like