Light

yangyangii / gan-tutorial Goto Github PK

View Code? Open in Web Editor NEW

370.0 4.0 106.0 3.53 MB

Simple Implementation of many GAN models with PyTorch.

Jupyter Notebook 100.00%

pytorch-gan gan gan-implementations dcgan conditional-gan wgan gan-pytorch pytorch gan-tutorial infogan

gan-tutorial's Introduction

GANs Tutorial

Very simple implementation of GANs, DCGANs, CGANs, WGANs, and etc. with PyTorch for various dataset (MNIST, CARS, CelebA).

You can run the code at Jupyter Notebook. And actually you can also run these codes by using Google Colab immediately (needed downloading some dataset)!

Sometimes ipynb files do not work in Github, please clone and run it in your server.

Requirements

python 3.6 (Anaconda)

pytorch 1.0.0 (updated from 0.4.0. If you want to use the previous version, then find previous commit.)

Implementation List

MNIST

Vanilla GAN

Conditional GANs

DCGAN

Wasserstain GAN

WGAN-gp

infoGAN

CARS (Stanford dataset)

you can download the dataset from https://ai.stanford.edu/~jkrause/cars/car_dataset.html

DCGAN

But I think the dataset is small and needs preprocessing a lot.

CelebA (aligned dataset)

you can download the dataset from http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html

BEGAN

R1GAN

Experimental Results

You can also see the samples at ipynbs.
After DCGAN, DCGAN with condition is a base model.
Trained 30 epochs respectively.

Vanilla GAN

Conditional GAN

DC GAN

WGAN-gp

infoGAN w/ walking code 1

infoGAN w/ walking code 2

BEGAN random samples (20 epochs)

BEGAN interpolation

GAN with R1 regularization random samples (20 epochs)

GAN with R1 regularization interpolation

Colab

https://colab.research.google.com
You can run it with GPU(K80) Runtime mode
Training Vanilla GAN takes only 10 minutes.

gan-tutorial's People

Contributors

Stargazers

Watchers

Forkers

ml-lab liaopeiyuan jinchoi96 charlesnord lmingyin fengzhangyin w4-jhyang aryeshdb hsiehkctaiwan53 deeplearningpk pzsuen megayeye eyuuo yogsin sixitingting yuekangdu lognat0704 kellsky mactenngre seaerchin kiniavinash inhyek juvu wengdunfang belye h-jia trungthanhnguyen0502 adrienju hircumg cjfdl-lab ryutian kuan-li scorpion0po jdc08161063 ezedinff scape1989 gheyret ducnh1022 davidpqc1231 skyorca arnaudyoh mohammadalm fluff18 dingkmc ragnariock hakandeymeci mastryukov1990 marque-antoine pink0763 ainilaha liangzhendong123 alexts1980 harveyla y-kanan bxyevonik pankii-09 yiyunzhao anhnt2407 xiaohehuanshu saiganeshrajakommuri zahra929 uri-barak yale1417 quantummole imadbouyazza nivedwho ogbaltaoglu sreyareddy15 fernandaecc chanakyayadav alexeusc4ever ivanlowllh asnsams iakujobi gitvicky chaoticbob hanzhouliu junjiehub eunsooyang jcarlosneto sunsern ds604 hulaba aslankims qwytt fsq2 malikbasit32 123hanyuchen ggh5454 firezl amrakibulhasan aanirudh07 wslucy szilviaujvary vejmin hkim09 uniqueonez 370025263 easy-forks anshshah3009

gan-tutorial's Issues

change size of image

Hi, thank you for making an amazing notebook that I can practice.

I would like to change the size of the image.

For example, in R1GAN and BEGAN, IMAGE_DIM is (64, 64, 3). However, I would like to change IMAGE_DIM as (512,512,3) .

I want my result of GAN also to become (512,512,3). Can you give me the instruction about how am I supposed to change the network?

Thank you

Should we freeze parameters in generator / discriminator when training the other?

Hi, thank you for this great repo, it helps me a lot.

I have a simple question about the training process. Should we freeze parameters in generator / discriminator when training the other?

It seems when running G_loss.backward(), parameters in discriminator will change with it?

not an issue: what is the function of the 10 in get_sample_image(G, n_noise)

def get_sample_image(G, n_noise):
    """
        save sample 100 images
    """
    z = torch.randn(10, n_noise).to(DEVICE)
    y_hat = G(z).view(10, 3, 28, 28).permute(0, 2, 3, 1) # (100, 28, 28)
    result = (y_hat.detach().cpu().numpy()+1)/2.
    return result

What is the deal with the 10? I understand the 100 is random noise as an input, but what about the 10? Thank you so much for your repo btw, very helpful to me.

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.