.
|-- Research
| |-- Code
| |-- TF-Level2-lectures\ references
| |-- CycleGAN
| |-- StarGAN
| |-- CGAN_Tutorial for MNIST
| |-- Paper\
| |-- Nara
| |-- PPT\
| |-- Jiyoon
| |-- Nara
| |-- Seungwon
|-- Verification
| |-- StyleGAN2-ada
| |-- code
| |-- result
|-- Test
| |-- code
| |-- result
| |-- 모델 설명
|-- submit\ Including Dockerfile
|-- Study_meeting\ Updated after every meeting
- Basic-GAN @jiyoon baek
- Conditional-GAN @haenara shin
- CycleGAN @seungwon song
- StarGAN @seungwon song
- StyleGAN2-ada @seungwon song
- Animal Transfiguration -> Attention-Gan,cycle gan @jiyoon baek I mistakenly deleted a cycle gan directory I made in my lab server computer by typing rm :( My initial attempts were to provide both corrupted images loaded from .ipynb file and datasets crawled from web portals but I really am sorry for what happened. Instead, I'll provide a tutorial code @lornatang uploaded. Thanks for providing amazing codes ! reference : cycle gan from lornatang
- Progressive Face Aging (PFA) GAN @haenara - Failed
.py 테스트코드 가이드
To convert image, we need target image that want to convert and W
that contains style information.
First, We extract W
from 2 sample images. One(sample_after
) is an image expressing a specific style(ex. smile, skin, age etc.),
The other(sample_before
) doesn't have that style (the more completely identical other features here, the better).
- input image :
sample before
,sample after
,target before
- output
W
: 'get_w.pt' (extracted Style by subtractingsample_before
fromsample_after
) - output image :
target after
.py 파일 Colab에서 실행하기 --> Example
!git clone https://github.com/sw-song/stylegan2-ada-pytorch.git
# wee need this package in colab
!pip install ninja
# move to the folder that we cloned
%cd stylegan2-ada-pytorch/
# run python command
!python conversion.py --sample_before s_b.png --sample_after s_a.png \
--target_before t_b.png --target_after t_a.png \
--network https://nvlabs-fi-cdn.nvidia.com/stylegan2-ada-pytorch/pretrained/afhqdog.pkl
- StyleGAN2-ada | 테스트 결과 | 시바견 @seungwon song
- Ada,Stylegan,Stylegan2 논문 설명 (모델 설명 📁) @jiyoon baek
- StyleGAN2-ada | 테스트 코드/결과 | 웰시코기, 시바견, 요키 @seungwon song ❎
- Style Transfer | 테스트 코드/결과 | 요키 @seungwon song ❎
- StyleGAN2-ada | 테스트 코드/결과 | 요키+body+background @seungwon song ❎
docker image build guide (assumed after the docker installation and launching)
- Build the Docker image:
docker build -t <docker image name> -f submit/Dockerfile .
- ex>
docker build -t docker_test -f submit/Dockerfile .
- Your working directory is the upper of
submit
folder
- ex>
- Generate the Docker container:
docker run --gpus all -it <docker image name>
- ex>
docker run --gpus all -it docker_test
- Your working directory is the upper of
submit
folder. - Once you successfully run, your working env will be changed to
root@something:/submit#
- ex>
- Run the
inference.py
:python inference.py
- Please wait and pray for the future.
- Check the ouput folder:
cd output/
, and thenls
to check the result (image, gif, and mp4 files)
6/ 7 (Mon)
| Roadmap & Strategy6/14 (Mon)
| Basic Research - GAN, ConditionalGAN6/19 (Sat)
| Basic Research - CycleGAN, StarGAN6/26 (Sat)
| Model Verification - StyleGAN2-ada✅ PFA-GAN❎ Cycle-GAN❎6/30 (Wed)
| Model Test - StyleGAN2-ada ❎ | Change our task to create a growing-up-video