may22es / gan_project Goto Github PK

.
|-- 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

• StyleGAN2-ada | 주피터 노트북 | 테스트 코드 @seungwon song
• StyleGAN2-ada | 파이썬 | 테스트 코드 @seungwon song

.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 subtracting sample_before from sample_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)

1. Build the Docker image: docker build -t <docker image name> -f submit/Dockerfile .
   1. ex> docker build -t docker_test -f submit/Dockerfile .
   2. Your working directory is the upper of submit folder
2. Generate the Docker container: docker run --gpus all -it <docker image name>
   1. ex> docker run --gpus all -it docker_test
   2. Your working directory is the upper of submit folder.
   3. Once you successfully run, your working env will be changed to root@something:/submit#
3. Run the inference.py: python inference.py
   1. Please wait and pray for the future.
4. Check the ouput folder: cd output/, and then ls to check the result (image, gif, and mp4 files)

• 6/ 7 (Mon) | Roadmap & Strategy
• 6/14 (Mon) | Basic Research - GAN, ConditionalGAN
• 6/19 (Sat) | Basic Research - CycleGAN, StarGAN
• 6/26 (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

1. GAN-ZOO
2. TF-GAN: 이활석님 TF-GAN repos.