search

ivanvovk / WaveGrad

Implementation of WaveGrad high-fidelity vocoder from Google Brain in PyTorch.
BSD 3-Clause "New" or "Revised" License
402 stars 56 forks source link

On CPU performance #2

Closed erogol closed 4 years ago

erogol commented 4 years ago

Hi thanks for your work and all the info in you README!

Did you measure the RTF on a CPU?

ivanvovk commented 4 years ago

Hi! I'm preparing it. Check it out in a few hours.

ivanvovk commented 4 years ago

I have updated RTF info in README. Surprisingly, but my 6-iteration model is a bit slower than the one reported in the paper. Perhaps I have an older model of CPU than authors used for their experiments. Note: you can always run RTF estimation for your own runtime environment on your own data or just random inputs using the code provided in notebooks/inference.ipynb notebook or in benchmark.py.