zlacker

>>Broken+(OP)
There seems to be a convincing debunking thread on Twitter, but I definitely don't have the chops to evaluate either claim:

https://twitter.com/Ethan_smith_20/status/175306260429219874...

>>mmastr+dr
Only took a quick glance. But it looks like a good debunking to me, especially the part where he refers to a section of the original paper clearly stating that the VAE has been trained with a 10e-6 weight factor on the KL divergence term.

I think what would happen if this problem was fixed is that the VAE would produce less appealing more blurry images. This is a classic problem with VAEs. So, more mathematically correct, but less visually appealing.