emiliendupont / wgan-gp Goto Github PK

View Code? Open in Web Editor NEW

239.0 5.0 50.0 18.57 MB

Pytorch implementation of Wasserstein GANs with Gradient Penalty

License: MIT License

Python 100.00%

wgan-gp pytorch wasserstein-gan gradient-penalty gan

wgan-gp's People

Contributors

Stargazers

Watchers

Forkers

zilinyan piegu johaupt sankin1770 bobby-he yangyang-nus-lv riknel memray stjordanis willshiao jeffshee dovedx mani-kt zyong812 xmpx jacksklar mbjoseph whynotbicycle ryuichi-tate chomd90 weihengxia0123 fedral eeyhsong archangel212 saeidnp dunkelhaus hendriktpl nahoon02 chenpaulyu taoxianpeng bunnyc1986 divdarshini z5614036 on1262 aramist op-real eaearth edwinb-ai shanespace amrakibulhasan lucasmspereira luciabarb supeking flazerain forkdzzx fabacha msxpwr vietld-itec jqjin123 guneet19

wgan-gp's Issues

Does this GAN also use EMD(Wasserstein distance)?

Hi!
I love this great job but I have difficulty reading the codes. Does this WGAN also use EMD i.e. the Wasserstein distance to approximate the real instances? I haven't find such a loss in loss_G or other losses, Thank you

License

Do you plan to add a license to this repo? Thanks!

Division Issue int typecast

The self.img_size[0]/16 needs to be typecasted to int:

`class Generator(nn.Module):
def init(self, img_size, latent_dim, dim):
super(Generator, self).init()

    self.dim = dim
    self.latent_dim = latent_dim
    self.img_size = img_size
    self.feature_sizes = (int(self.img_size[0] / 16), int(self.img_size[1] / 16))`

Training minist dataset problem

I can't Reappearance that Results like issue 4.
And I find it that this gif what my model generated 6.60Mb,
But in this code example gif folder,it is 8.22 Mb.
What make this difference?
I hope author to anser this question,thank you very much!

Results not reproducible

Thanks for sharing this repo.

Running the code (main.py, 200 epochs, parameters as mentioned in readme, only fixing the the issue mentioned in #1), I couldn't reproduce the results shown in https://github.com/EmilienDupont/wgan-gp/blob/master/gifs/mnist_200_epochs.gif. Instead, the generated samples look as follows:

Any hints on what would need to be changed to make the training successful?

One step towards reproducibility would be to fix random seeds in the beginning of main.py, e.g. with the following code:

import random
import numpy as np
import torch
random_seed = 42
random.seed(random_seed)
np.random.seed(random_seed)
torch.manual_seed(random_seed)
torch.cuda.manual_seed_all(random_seed)

emiliendupont / wgan-gp Goto Github PK

wgan-gp's People

Contributors

Stargazers

Watchers

Forkers

wgan-gp's Issues

Does this GAN also use EMD(Wasserstein distance)?

License

Division Issue int typecast

Training minist dataset problem

Results not reproducible

Recommend Projects

React

Vue.js

Typescript

TensorFlow

Django

Laravel

D3

Recommend Topics

javascript

web

server

Machine learning

Visualization

Game

Recommend Org

Facebook

Microsoft

Google

Alibaba

D3

Tencent