Papers of the day   All papers

GANSpace: Discovering Interpretable GAN Controls

Comments

roadrunner01: GANSpace: Discovering Interpretable GAN Controls pdf: https://arxiv.org/pdf/2004.02546.pdf abs: https://arxiv.org/abs/2004.02546 video: https://youtu.be/jdTICDa_eAI github: https://github.com/harskish/ganspace https://t.co/OsRIpdEpjz

3 replies, 383 likes


Aaron Hertzmann: Check out our new paper on discovering of controls for making images with GANs, without any supervision (other than optionally naming the controls afterward), by Erik Härkönen, with @jaakkolehtinen, @sylvain_paris, and myself. https://arxiv.org/abs/2004.02546 https://t.co/VX1unSC9UW

5 replies, 376 likes


Jaakko Lehtinen: How to poke pretrained GANs and get them to do your bidding, my @AaltoCS student Erik’s internship project with @AaronHertzmann and @sylvain_paris. Super happy for this new collaboration!

1 replies, 32 likes


Fabian Offert: This is absolutely brilliant. So simple and effective. Congrats! #GAN #XAI

0 replies, 2 likes


arXiv CS-CV: GANSpace: Discovering Interpretable GAN Controls http://arxiv.org/abs/2004.02546

0 replies, 1 likes


AaltoMediaAI: GANs getting more controllable and usable for image synthesis

0 replies, 1 likes


Content

Found on Apr 07 2020 at https://arxiv.org/pdf/2004.02546.pdf

PDF content of a computer science paper: GANSpace: Discovering Interpretable GAN Controls