deepspeech-cleaner icon indicating copy to clipboard operation
deepspeech-cleaner copied to clipboard

Not working with CommonVoice data

Open smddzcy opened this issue 4 years ago • 0 comments

Hey! First of all, thanks for this great little helper.

I think CommonVoice recently changed its download format and this library is not working with that anymore. I had to download the files manually, convert .tsv files to .csv (yes, CV gives you .tsv now), change a bit of code in downloader.py to make it work. Just letting you know.

Small Perl script for .tsv -> .csv conversion = perl -lpe 's/"/""/g; s/^|$/"/g; s/\t/","/g' < input.tsv > output.csv Also I changed line 193 in downloader.py, ...['/cv-other-train.csv to ...['/train.tsv.

smddzcy avatar Apr 12 '20 17:04 smddzcy