32 ms·
I’d be interested in reading your formal write up. It’d be interesting to see the issue as you understand it.
by sjg007 6y ago
I’d be interested in reading your formal write up. It’d be interesting to see the issue as you understand it.
- sillysaurusx 6y agoSure thing! https://github.com/google/compare_gan/issues/54 https://github.com/google/compare_gan/issues/54 It’s not much of a writeup. It’s basically saying, hey, this is zero when it should be one. The results were dramatic. It went from blobs to replicating the biggan paper almost perfectly. I think we’re at a FID of 11 or so on imagenet. Here's a screenshot I just pulled from our current run: https://i.imgur.com/k1RuWEG.png https://i.imgur.com/k1RuWEG.png Stole a year of my life to track it down. But it was a puzzle I couldn’t put down. It haunted my dreams. I was tossing and turning like, but why won’t it work... why won’t it work...