search

lucidrains / audiolm-pytorch

Implementation of AudioLM, a SOTA Language Modeling Approach to Audio Generation out of Google Research, in Pytorch
MIT License
2.33k stars 249 forks source link

Training curves for discriminator losses? #165

Closed turian closed 1 year ago

turian commented 1 year ago

I am curious if someone can share what the training curves look like for the discriminator losses?

@djqualia perhaps?

I am wondering if on my particular corpus, they are dropping fast enough.

lucidrains commented 1 year ago

think this can be moved to discussions