Add training_step method to models

CAIIVS / chuchichaestli

Where you find all the state-of-the-art cooking utensils (salt, pepper, gradient descent... the usual).

GNU General Public License v3.0

4 stars 0 forks source link

Open phdenzel opened 3 days ago

phdenzel commented 3 days ago

Different models require different (default) trainings, e.g.

VAE: L2 + KL
VAE for latent diffusion: L2 + KL + LPIPS + GANLoss (adaptive weighting as in Eq. 7 from https://arxiv.org/pdf/2012.09841)