๐ข Exciting News! Our paper on StyleDomain for One-shot and Few-shot Domain Adaptation, accepted to ICCV 2023, is out! ๐๐ฅ
๐ Paper Link: https://t.co/XTvjw9jhEV
๐ Source Code: https://t.co/M4die1h6DC
#ICCV2023#GANs#StyleDomain#DomainAdaptation
1/N ๐งต
๐ฆ Get Your Hands Dirty!
Excited to try it out? You can find the source code for StyleDomain on our GitHub repository: https://t.co/MZctTUB4PI. Experiment, adapt, and explore the potential of domain adaptation with StyleGAN.
6/N
You can play with our models here: https://t.co/aIv9M4rLM0
We set up the colab notebook for you that you can easily run and get all results of the paper!
This work would not be possible without amazing collaborators Vadim Titov and Dmitry Vetrov!
N/N
Excited to share our new #NeurIPS2022 paper! We significantly reduce the number of training parameters for adapting StyleGAN2 to new domains: from 30 million to 6 thousand!
Paper: https://t.co/anX6PiIStv
Code: https://t.co/aIv9M4rLM0
1/N
Experimentally we show that our technique has comparable quality with the existing SOTA domain adaptation methods while utilizing by several orders less number of training parameters! We justify it both quantitatively and qualitatively.
4/N
Our new paper, Truncated Quantile Critics, improves SOTA on MuJoCo by 20-30% ! With TF and PT code.
Credit to @brickerino@shvechikov_p@AlexGrishin_ and Dmitry Vetrov.
Video: https://t.co/o0zvcuHa8X
Project page: https://t.co/ppZi3Tc1VJ
Paper: https://t.co/XodndN2enO