Skip to content

HaitaoWuTJU/Deep-Feature-Consistent-VAE

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Deep Feature Consistent VAE With PaddlePaddle

1.Download CelebA and Pretrained VGG19

Download CelebA @AI Studio CelebA and put data at ./dataset

Download pretrained VGG19_bn @PaddlePaddle (password: 25il) and put data at ./paddle_dfcvae/architectures/pretrained

then

sh preparedata.sh

2.Train paddle dfcvae

cd ./paddle_dfcvae
python trainer.py

chose vgg layer feature

vim  extract_features @ ./paddle_dfcvae/architectures/dfcvae.py

3.Generated imgs

python generatedImg.py

4.Reconstruct

vae123 Reconstruct

vae345 Reconstruct

vae123 extract the low features,clearer than vae345

5.Generated

Generated Generated

6.Logs , Models and More Imgs

BaiduYun (password: 9vs9)

19967 Reconstruct pics(2x10)

100 Generated pics(10x1)

7.Measure

IS:

mean std
IS 1.1636 0.0221
cifar10 1.1698 0.0355

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published