Hi Alex,
thanks for your work.
I was wondering if I could blow this architecture up.
I roughly think of something like this:
1024² as outputshape.
512(?)-size latentspace. I am very unsure about this, but have seen that you have used different sizes here for your different datasets. What approach did you use to estimate this?
Around 2000-3000 training examples.
Do you have any experience with this or can tell me already why this cannot work?
Hi Alex, thanks for your work. I was wondering if I could blow this architecture up.
I roughly think of something like this: 1024² as outputshape. 512(?)-size latentspace. I am very unsure about this, but have seen that you have used different sizes here for your different datasets. What approach did you use to estimate this? Around 2000-3000 training examples.
Do you have any experience with this or can tell me already why this cannot work?
Cheers, Joscha