Blog Archive

Thursday, June 13, 2019

how to download voxceleb data automatically and prepare for kaldi experiment

Part 1: folder structure preparation:
suppose voxceleb 1 & 2 need to be downloaded into: voxceleb1 & voxceleb2
create dev, test under each directory and put all dev files into dev, test into test
i.e, the final folder directory should be:
mkdir -p voxceleb1/dev voxceleb1/test
mkdir -p voxceleb2/dev voxceleb2/test


Part 2: data download
wget --user=XX --password=YY http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1a/vox2_test_aac.zip

You can follow the above example on both voxceleb1 and voxceleb2
Note you need fill this form to request a password.

https://github.com/kaldi-asr/kaldi/tree/master/egs/voxceleb/v1


Reference:
http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html
http://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox2.html
https://askubuntu.com/questions/29079/how-do-i-provide-a-username-and-password-to-wget

No comments:

Post a Comment