11 comments

[ 2.4 ms ] story [ 33.7 ms ] thread
This is pretty interesting and I'd be curious to understand the technique and sources behind it. I figure it is some kind of machine learning going on but I assume that there is a number of these that have been pre-generated and this is kind of a random slideshow. They look like real people.
It is based on https://arxiv.org/pdf/1812.04948.pdf

Images can easily be generated on the fly. Inference time is ~100ms.

A similar site, based on the same algortithm, was extensively discussed a few days ago: https://news.ycombinator.com/item?id=19144280

What's different about the site you submitted?

Thanks! I missed that one - seems to do the same :-)
Oh yeah... You "missed" that one....
It would be useful to add some parameters, for instance allowing to choose an age range.
good suggestion. Looking into latent vector manipulation now.
Would be interesting to see the results of that.
You must have been sitting on this idea for decades seeing as how you registered the domain on 2-16-2019. And you have the guts to say this isn't a blatant copycat? Sheesh.
Why this negativity? It's not harming anyone. Just fun to play with the new StyleGAN network which is only out since a few weeks.