maua-stylegan2
This is the repo for my experiments with StyleGAN2. There are many like it, but this one is mine. Contains code for the paper Audio-reactive Latent Interpolations with StyleGAN. (by JCBrouwer)
stylegan2-surgery
StyleGAN2 fork with scripts and convenience modifications for creative media synthesis (by aydao)
maua-stylegan2 | stylegan2-surgery | |
---|---|---|
2 | 3 | |
179 | 143 | |
- | - | |
0.0 | 0.0 | |
almost 3 years ago | over 3 years ago | |
Python | Jupyter Notebook | |
GNU General Public License v3.0 or later | GNU General Public License v3.0 or later |
The number of mentions indicates the total number of mentions that we've tracked plus the number of user suggested alternatives.
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.
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.
maua-stylegan2
Posts with mentions or reviews of maua-stylegan2.
We have used some of these posts to build our list of alternatives
and similar projects. The last one was on 2022-10-04.
-
I'm stumped with installing PyTorch.
Originally I wanted to run https://github.com/JCBrouwer/maua-stylegan2. I was trying to run the convert_weight.py but it resulted in shape mismatch errors in torch torch.Size([1, 512, 4, 4]) vs torch.Size([1]), so I tried the version here https://github.com/rosinality/stylegan2-pytorch/blob/master/convert_weight.py and the result was the same.
-
Pretrained 1792x1024 StyleGAN2 model
You don't even need to really do model surgery. All the convolutions will accept arbitrary dimensions. You can just use network bending padding operations to get any output size you like Vadim Epstein's repo does something slightly different which let's you even use different latents per section: https://github.com/eps696/stylegan2ada Or mine which has the simpler, single latent version https://github.com/JCBrouwer/maua-stylegan2 Or for training, then all you have to do is change the size of your constant layer Or just graft on some more upsamples Either way though, there's not too much point to training at weird rectangular resolutions. You'll get pretty much identical results by just forcefully resizing to a square and then stretching the generated versions back out to square Unless you've got a ridiculous amount of VRAM, larger models don't really make too much sense either. Especially because it'll be hard to find 10k images at such a big resolution
stylegan2-surgery
Posts with mentions or reviews of stylegan2-surgery.
We have used some of these posts to build our list of alternatives
and similar projects. The last one was on 2022-09-13.
-
AI-generated Formidable portraits
Thanks! From scratch it takes from days to a few weeks with a powerful GPU, but if one transfer-learns from an existing model (like ThisWaifuDoesNotExistv3) it can be done in a few hours-days and some additional hours for finishing touches (lowering learning rate, clearing dataset with the discriminator). Most of the active work goes into collecting and preparing the data, with some spent with manual evaluation for best model checkpoints. To my understanding there are very few attempts to generate more than an anime head with these, and I'm not sure how much stylegan would need to be modified for that. There is this project which I have heard might work with full-body generation, but I have no experience with it.
- Pretrained 1792x1024 StyleGAN2 model
-
Unable to decode bytes as JPEG, PNG, GIF, or BMP
I get this error using this repo https://github.com/aydao/stylegan2-surgery . I checked all Images in the dataset, their not corrupted and all have the right file format
What are some alternatives?
When comparing maua-stylegan2 and stylegan2-surgery you can also consider the following projects:
stylegan2ada - StyleGAN2-ada for practice
stylegan2-pytorch - Implementation of Analyzing and Improving the Image Quality of StyleGAN (StyleGAN 2) in PyTorch
stylegan3 - Official PyTorch implementation of StyleGAN3