Hi, is anyone trying to apply this?
1 Like
I am currently doing it, if I have any sucess I will let you know
2 Likes
I have implemented the three steps mentioned in the paper
as Time Warp, Frequency Mask followed by Time mask. Though the spectrogram looks fine, but not sure how much difference it will have. Currently I am validating my implementation.
I ended up with a lot of problems, could you share your implementation?
Thanks a lot!
There’s a PR here: https://github.com/mozilla/DeepSpeech/pull/2090