News for hackers. Stuff that matters.

A comment on WaveNet implementation in Keras. Parent comment

Re:WaveNet implementation in Keras by basve on Thursday September 15, 2016 @06:58AM (#12504916)
And I should add that this was measured using a downsized model (just two blocks of dilated convolutions and a sampling rate of 4khz). Deepmind's paper does not report how many stacks are used to generate the samples, but I assume it's quite a bit more.
[ Parent | on HN ]
Re:WaveNet implementation in Keras by unlikelymordant on Thursday September 15, 2016 @07:08AM (#12504972)
They (deepmind) reported it took 90 minutes of processing to generate 1s of speech via tweet. Hopefully this comes down in the future.
[ Parent | on HN ]
Re:WaveNet implementation in Keras by espadrine on Thursday September 15, 2016 @09:30AM (#12505939)
Do you have a link?

This implementation says: “A Tesla K80 needs around ~4 minutes for generating a second of audio at a sampling rate of 4000hz”, which is significantly faster.

[ Parent | on HN ]
Re:WaveNet implementation in Keras by basve on Thursday September 15, 2016 @09:46AM (#12506044)
90 minutes for 1s of audio was reported by someone from Google on twitter, but the tweet has been deleted. I've clarified in the readme that my measurements are for a much lighter/smaller model than Deepmind's :).
[ Parent | on HN ]