I am using PyTorch 1.9.0.dev20210512+cu111 version to train the Gaussian example. Without any modification but changing demo.json to the gaussian.json (also project name as well), I received the warning Using a non-full backward hook when the forward contains multiple autograd Nodes
Just wondering does this warning have any impact on the trained model?
Overall, really appreciate for the hard work putting on this project. Even at 5 hrs training audio I can get something reasonably satisfied. Great Thanks!
I am using PyTorch
1.9.0.dev20210512+cu111
version to train the Gaussian example. Without any modification but changingdemo.json
to thegaussian.json
(also project name as well), I received the warningUsing a non-full backward hook when the forward contains multiple autograd Nodes
Just wondering does this warning have any impact on the trained model?
Overall, really appreciate for the hard work putting on this project. Even at 5 hrs training audio I can get something reasonably satisfied. Great Thanks!