Bomze

Face Depixelizer Given a low-resolution input image, model generates high-resolution images that are perceptually realistic and downscale correctly. GitHub: https://github.com/tg-bomze/Face-Depixelizer … Colab: https://colab.research.google.com/github/tg-bomze/Face-Depixelizer/blob/master/Face_Depixelizer_Eng.ipynb … P.S. Colab is based on the https://github.com/adamian98/pulse pic.twitter.com/lplP75yLha

Jukebox: A Generative Model for Music This combined model at scale can generate high-fidelity and diverse songs with coherence up to multiple minutes. GitHub: https://github.com/tg-bomze/Jukebox_Colab … Colab: https://colab.research.google.com/github/tg-bomze/Jukebox_Colab/blob/master/Jukebox_Eng.ipynb … P.S. Colab is based on the https://github.com/openai/jukebox pic.twitter.com/EHDyAli2qj

#MonaLisa Experimentpic.twitter.com/IIyBcFh1U9

Timbre Transfer (Voice2Accompaniment) This algorithm enables high-fidelity audio synthesis without the need for large autoregressive models. GitHub: https://github.com/tg-bomze/TimbreTransfer … Colab: https://colab.research.google.com/github/tg-bomze/TimbreTransfer/blob/master/TimbreTransfer_Eng.ipynb … P.S. Colab is based on the https://github.com/magenta/ddsp pic.twitter.com/u4cMnWliw3

Stylizer Video to Drawing. This project generates artistic portrait drawings from face video frames using a GAN-based model. GitHub: https://github.com/tg-bomze/Video2Drawing … Colab: https://colab.research.google.com/github/tg-bomze/Video2Drawing/blob/master/Video2Drawing_(Eng).ipynb … P.S. Colab is based on the https://github.com/yiranran/Unpaired-Portrait-Drawing ….pic.twitter.com/K7HTmfisgp

Video to Anime Converter. This is Colab, which uses the GAN framwork to transform real-world video frames into anime. GitHub: https://github.com/tg-bomze/Video2Anime … Colab: https://colab.research.google.com/github/tg-bomze/Video2Anime/blob/master/AnimeGan_Video_(Eng).ipynb … P.S. Colab is based on the https://github.com/TachibanaYoshino/AnimeGAN ….pic.twitter.com/hxsyktK5Hl

Face Image Motion Model. I rebuilt the popular model, adding interpolation and upscale frames. Detailed DoomGuy with Remini and StyleGan2 (from my build). GitHub: https://github.com/tg-bomze/Face-Image-Motion-Model … Colab: https://colab.research.google.com/github/tg-bomze/Face-Image-Motion-Model/blob/master/Face_Image_Motion_Model_(Photo_2_Video)_Eng.ipynb … P.S. Colab is based on the https://github.com/AliaksandrSiarohin/first-order-model ….pic.twitter.com/iFqIIWWBfh

ENTAR v.1.0 - Ensemble of Neural Tools for Animations Restoration Ensemble includes: #DeOldify, #DeepRemaster, #VFIASC, #FastSRGAN, #3DKenBurns. Thank everyone who made these cool neural networks! Have fun: https://colab.research.google.com/github/tg-bomze/ENTAR/blob/master/ENTAR_Eng.ipynb …pic.twitter.com/w46tjjt9rL

Restoration of Black&White old video with Charlie Chaplin. List of neural networks: #DeOldify, #FastSRGAN, #DeepRemaster, #DAINpic.twitter.com/IVQW37WBNk

The last video about restoration and rejuvenation on this account. Similar historical transformations I will be published here: https://www.behance.net/malimonoffart . On Twitter, we will continue to dive deep into neural networks. #Mozart #StyleGAN2pic.twitter.com/PdedsEZpUy

#Pushkin #StyleGAN2pic.twitter.com/b4toQnQM2G

#TheBirthOfVenus #StyleGAN2pic.twitter.com/rMZ9xr5ANh

#MonaLisa #StyleGANpic.twitter.com/p7l9vcfWWm