Where is the optimized adaptive channel importance scores?

[songquanpeng]

Hi, thx for your great work! But I have a question about the implementation. In the paper, the objective function is:

So it looks like that we should get the optimized latent codes and the adaptive channel importance scores.

But in your implementation, there are only the latent codes are optimized:

# Invert
latent_estimates, history = inversion.invert(generator, y_gt, loss, batch_size=1, video=args.video)

https://github.com/genforce/mganprior/blob/master/multi_code_inversion.py#L36

I wonder if there is something wrong with my understanding.

Thx in advance!

[songquanpeng]

Got it