prism
pytorch-AdaIN
prism | pytorch-AdaIN | |
---|---|---|
5 | 1 | |
101 | 1,001 | |
- | - | |
0.0 | 0.0 | |
almost 2 years ago | 4 months ago | |
Python | Python | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
prism
-
Stable Diffusion might be the holy grail.
I'll point you towards Prism, then. As far as style transfer goes, it's the best I've found so far. Extremely simple to run in colab and offline, plus the results are quite good for what it's doing. I used it a ton before I started running Face Portrait v2 offline as a general "art" filter for things I'm working on. It does stylization and some facial morphing in one pass, which adds a lot more personality than style transfer alone. You can also use it on anything, not just faces.
-
Best High Resolution Style Transfer
My best results are with https://github.com/moritztng/prism, but this only gets me to ~4MP max.
-
How is it Possible To Create High Res Images With Low Res Network?
Naive question here. I notice a lot of style transfer libraries (e.g. https://github.com/moritztng/prism) can create high res images (2048x2048) or bigger.
-
The Apple, Me, AI-Assisted Digital Collage, 2021
This is a mix of original photography, Prism, and some quick finalizations in Photoshop. The goal was to create something of an ironic still life.
-
I built a shop based on style transfer where you can buy personal artworks as canvas prints
Code: https://github.com/moritztng/prism
pytorch-AdaIN
-
[P] Can't finish my master's thesis. What to do?
Thanks for the comment. I am using the 3d views directly. I tried to use some form of style transfer (https://github.com/naoto0804/pytorch-AdaIN) from real to synth data, but the result was'nt that appealing, so I moved on. I guess that it could be due to the fact that generally the real data that I have has low resolution (120x40 tipically). I ll have a check on the model that you suggested though, seems interesting.
What are some alternatives?
style-transfer-pytorch - Neural style transfer in PyTorch.
pytorch-neural-style-transfer - Reconstruction of the original paper on neural style transfer (Gatys et al.). I've additionally included reconstruction scripts which allow you to reconstruct only the content or the style of the image - for better understanding of how NST works.
style-transfer-app - An asynchronous dual application (web + Telegram bot) for stylization images.
contrastive-unpaired-translation - Contrastive unpaired image-to-image translation, faster and lighter training than cyclegan (ECCV 2020, in PyTorch)
animegan2-pytorch - PyTorch implementation of AnimeGANv2
neural-style-pt - PyTorch implementation of neural style transfer algorithm
stylized-neural-painting - Official Pytorch implementation of the preprint paper "Stylized Neural Painting", in CVPR 2021.