Comments (5)
Hi @maxrumi ,
A few questions:
- Are you still training with no labeled data?
- How do the results look on the train set?
Since it appears that you are using a different StyleGAN for a different task, using our toonify parameter settings as is probably will not lead to optimal results right from the start. You should play around a bit more with the parameters settings.
In any case, please provide a few more details about how you are training and maybe a few more results on the train set and I will hopefully be able to provide you with more concrete recommendations.
from pixel2style2pixel.
1,Are you still training with no labeled data?
Answer: Our psp model trained on paired data, the source domain dataset is ffhq dataset, and the target domain dataset is transfered ffhq dataset(by using other style transfer method), and our stylegan2 is trained on transfered ffhq dataset too.
2,How do the results look on the train set?
Answer: The picture showed on page top is the psp results on train data.
So, we just want to know how to change the parameter setting to make the psp results have same shape and texture just like the input data, and just transfer the input to some kind of cartoon style(like cycle gan).
Can you give us some parameter setting sample, suppose you just want to transfer the input image to cartoon style and don't want to change it's shape and texture.
We want to know how to change these parameters setting, thanks a lot:
--lpips_lambda=0.08
--l2_lambda=0.001
--lpips_lambda_crop=0.8
--l2_lambda_crop=0.01
--id_lambda=1
--w_norm_lambda=0.005
from pixel2style2pixel.
While I don't know exactly what loss settings will work best for you, I can try giving you some recommendations:
- Please note that it seems like you copied these parameter settings from the frontalization task, which is not really what you want in your task. You should consider starting with the parameters used in the toonification task (the lambda values are provided in the README).
- With that said, our toonification training is done with no paired data, so you will most likely need to play around some more with the loss settings, but this should be a good starting point.
Does this help answer your question?
from pixel2style2pixel.
Closing due to inactivity
from pixel2style2pixel.
我们的psp模型是在paired data上训练的,源域数据集是ffhq数据集,目标域数据集是transferred ffhq数据集(通过其他风格迁移方法),我们的stylegan2也是在transferred ffhq数据集上训练的。
@maxrumi
Hello, how did you generate your paired data set? Is this effect obtained by directly using the psp network model to train the paired data? Can you tell me about the implementation process?
from pixel2style2pixel.
Related Issues (20)
- Use the pre-trained model for training HOT 2
- do a huggingface demo HOT 1
- How to train on paired data? HOT 1
- latent image editing HOT 1
- About celebs_seg_to_face HOT 1
- multiple GPUs HOT 4
- Using my own pretrained model from vanilla StyleGAN2 HOT 3
- Is it possible video2anime train? like this project? HOT 3
- A problem about how to get diverse images? HOT 1
- the output image become brown HOT 3
- Need help with running the code on CPU HOT 1
- loading pretrained weights, STACK_GLOBAL requires str HOT 1
- How to train pSp on Z+ space? HOT 2
- Is it possible to create output images in profile (side-on) perspective using sketch to face? HOT 1
- Single channel input with Moco loss not working. HOT 2
- Some error reporting problems encountered during operation HOT 1
- How to use other identity loss rather than Arcface or Moco HOT 2
- How to use psp for Face beautification
- loss jump problem HOT 1
- Retraining
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from pixel2style2pixel.