#LatentFonts as the next iteration of #VariableFonts? Tried #StyleGAN on a 50k fonts dataset collected by @fulhack - Now I'm stuck watching this loop

8:39 AM · Feb 13, 2019

9
230
20
888
35,967
A higher-res export highlights something I didn't know I had: a favorite letter and that's lowercase y
2
8
0
47
2,219
Wow, would love to see more varieties
0
0
0
1
ML definitely renders variable fonts less interesting, but one should aknowledge the fact that these are not actual fonts, merely collections of bitmaps and that the crude resolution helps to hide their flaws. But definitely awesome to look at! cc: @hardmaru
2
1
0
10
Definitely. Despite the progress, I still believe that we need to model vector elements for these models to be more useful to designers.
1
1
0
14
Hi, nice work! did you use any label in the training procedure?
1
0
0
2
Thanks! No label, just raw images so far
1
0
0
2