martinarjovsky/WassersteinGAN issues and pull requests

#81 - I cannot find the calculating or estimating of wasserstein distance!

Issue - State: closed - Opened by deahhh over 2 years ago - 3 comments

#80 - Generator update

Issue - State: closed - Opened by JurijsNazarovs almost 3 years ago - 1 comment

#79 - Why did not tell the label to the discriminator

Issue - State: open - Opened by chenyuhaosuai almost 4 years ago

#78 - How can I use a loss as the stopping criteria in Wasserstein GAN?

Issue - State: open - Opened by RamziFsm almost 4 years ago

#77 - Interpreting Generator and Critic loss

Issue - State: open - Opened by KhrystynaFaryna over 4 years ago - 1 comment

#76 - some problem when running the WassersteinGAN

Issue - State: open - Opened by zsyai799 over 4 years ago - 1 comment

#75 - should the gamma and beta on batchnormalization layer be clipped?

Issue - State: open - Opened by dawnleft about 5 years ago

#74 - No sigmoid activation for G on MLP?

Issue - State: open - Opened by druzkaya about 5 years ago

#73 - No convergence in onw dataset

Issue - State: open - Opened by ghost about 5 years ago

#72 - Inconsistent loss function from the paper?

Issue - State: open - Opened by realcrane over 5 years ago - 8 comments

#71 - Generate images script

Pull Request - State: closed - Opened by kopytjuk over 5 years ago - 1 comment

#70 - Remove dots from module names to execute the code with pytorch 1.0.0.

Pull Request - State: closed - Opened by kopytjuk over 5 years ago - 1 comment

#69 - Bugfix to run the code on windows machines.

Pull Request - State: closed - Opened by kopytjuk over 5 years ago - 1 comment

#68 - module name can\'t contain "."

Issue - State: open - Opened by IlSLY almost 6 years ago - 1 comment

#67 - Results cannot be reproduced.

Issue - State: closed - Opened by zeakey almost 6 years ago - 2 comments

#66 - The parameter ‘db_path' of LSUN setting in 'main.py' should be changed to 'root'

Issue - State: open - Opened by whuCanon almost 6 years ago

#65 - change readme

Pull Request - State: closed - Opened by zyoohv almost 6 years ago

#64 - Results on cifar10 very bad even if trained for over 1000 epochs

Issue - State: closed - Opened by FANG-Xiaolin almost 6 years ago - 1 comment

#63 - how to train a 256128 image dataset and output 256128 result？

Issue - State: open - Opened by GuiQuLaiXi almost 6 years ago - 3 comments

#62 - cifar10 result not good as expect !

Issue - State: closed - Opened by zyoohv almost 6 years ago - 7 comments

#61 - Problems with the optimization of loss.

Issue - State: open - Opened by XJTUDog about 6 years ago - 4 comments

#60 - Why have a tensor of 1 or -1 in loss.backward()?

Issue - State: closed - Opened by arpan-dhatt about 6 years ago - 1 comment

#59 - Where can I find bibtex of Wasserstein GAN and related works?

Issue - State: closed - Opened by ghost over 6 years ago - 1 comment

#58 - After training the model, how to generate Test samples using the generator?

Issue - State: open - Opened by rohitsaluja22 over 6 years ago

#57 - Why is Adam solver an option?

Issue - State: open - Opened by meder411 over 6 years ago

#56 - Reference parameters for large resolution training

Issue - State: open - Opened by snakers4 over 6 years ago - 2 comments

#55 - Add Tensorboard logging for scalars and images, add FastImageFolder for large folders

Pull Request - State: open - Opened by snakers4 over 6 years ago

#54 - Bigger batchsize?

Issue - State: open - Opened by qiuxin2012 over 6 years ago - 1 comment

#53 - Question about the biases in the Convolutional Layers

Issue - State: open - Opened by tensorfreitas over 6 years ago

#52 - Update dcgan.py

Pull Request - State: open - Opened by innerlee almost 7 years ago

#51 - Update main.py

Pull Request - State: open - Opened by innerlee almost 7 years ago

#50 - Clip too early

Issue - State: open - Opened by innerlee about 7 years ago - 1 comment

#49 - confirm

Issue - State: open - Opened by songhaohaohao about 7 years ago - 3 comments

#48 - lmdb as a dependency

Issue - State: open - Opened by ypuzikov about 7 years ago - 2 comments

#47 - How to setup the [lsun-train-folder]?: python main.py --dataset lsun --dataroot [lsun-train-folder] --cuda

Issue - State: open - Opened by fenfeibani about 7 years ago - 1 comment

#46 - D Loss did not decrease

Issue - State: open - Opened by tgong2002 about 7 years ago

#45 - some problem when run WassersteinGAN

Issue - State: open - Opened by wangchao12 about 7 years ago

#44 - How to make some change for mnist dataset?

Issue - State: closed - Opened by JiahaoYao about 7 years ago - 2 comments

#43 - some problem run program in my tensorflow

Issue - State: open - Opened by wangchao12 about 7 years ago - 1 comment

#42 - Why there is no netG.parameters.requires_grad = False when updating netD?

Issue - State: closed - Opened by fungtion over 7 years ago - 2 comments

#41 - Question about 'one' and 'mone'

Issue - State: closed - Opened by sz128 over 7 years ago - 4 comments

#40 - Using batch normalization in the generator network leads to mode collapse

Issue - State: closed - Opened by yliu120 over 7 years ago - 1 comment

#39 - Fixing wrong variable name

Pull Request - State: closed - Opened by keskarnitish over 7 years ago - 1 comment

#38 - compatible with latest pytorch on nn.parallel.data_parallel

Pull Request - State: closed - Opened by eriche2016 over 7 years ago - 1 comment

#37 - compitible with latest pytorch such that nn.parallel.data_parallel w…

Pull Request - State: closed - Opened by eriche2016 over 7 years ago

#36 - pytorch code for Improved Training of Wasserstein GANs

Issue - State: closed - Opened by aosokin over 7 years ago - 18 comments

#35 - de-normalize images for displaying

Pull Request - State: closed - Opened by eriche2016 over 7 years ago - 1 comment

#34 - de-normalize the images for displaying

Pull Request - State: closed - Opened by eriche2016 over 7 years ago

#33 - convert the real images and generated images back to original format to display???

Issue - State: closed - Opened by eriche2016 over 7 years ago - 3 comments

#32 - Question on clamp weights

Issue - State: closed - Opened by hengyuan-hu over 7 years ago - 1 comment

#31 - Questions about loss curve

Issue - State: closed - Opened by ghost over 7 years ago - 1 comment

#30 - ask for code

Issue - State: closed - Opened by ghost over 7 years ago - 1 comment

#29 - Generator non linearity missing?

Issue - State: closed - Opened by dmarnerides over 7 years ago - 1 comment

#28 - Have you trained in celebA dataset?

Issue - State: closed - Opened by zhangqianhui over 7 years ago - 4 comments

#27 - Quetions about the D out

Issue - State: closed - Opened by yichuan9527 over 7 years ago - 3 comments

#26 - Why is data always normalized w/ hardcoded mean/std = 0.5

Issue - State: closed - Opened by mjdietzx over 7 years ago - 1 comment

#25 - set noisev to Volatile to freeze netG when updating netD.

Pull Request - State: closed - Opened by chenyuntc over 7 years ago - 1 comment

#24 - parameterize input image channels

Pull Request - State: closed - Opened by edgarriba over 7 years ago - 1 comment

#23 - show epoch progress correctly

Pull Request - State: closed - Opened by edgarriba over 7 years ago - 1 comment

#22 - Error in python: double free or corruption (out)

Issue - State: closed - Opened by zhoulw13 over 7 years ago - 1 comment

#21 - Questions about the D_loss

Issue - State: closed - Opened by yichuan9527 over 7 years ago - 2 comments

#20 - resize problem

Issue - State: closed - Opened by yichuan9527 over 7 years ago - 3 comments

#19 - fix some resize related bugs

Pull Request - State: closed - Opened by elPistolero over 7 years ago - 1 comment

#18 - Batchnorm Scaling Factor is Clamped Near Zero

Issue - State: closed - Opened by whuffman over 7 years ago - 7 comments

#17 - is the estimate of Wasserstein distance negative?

Issue - State: closed - Opened by aosokin over 7 years ago - 3 comments

#16 - Conditional WGAN

Issue - State: closed - Opened by aosokin over 7 years ago - 23 comments

#15 - inputv update

Issue - State: closed - Opened by elPistolero over 7 years ago - 1 comment

#14 - Mode collapse in Conditional GAN Sequence (text) generation

Issue - State: closed - Opened by rafaelvalle over 7 years ago - 17 comments

#13 - IOError: image file is truncated

Issue - State: closed - Opened by xiesiyuan over 7 years ago - 3 comments

#12 - Fixed noise and conditions for Wasserstein GAN training

Issue - State: closed - Opened by rafaelvalle over 7 years ago - 1 comment

#11 - How to compute the WGAN loss gradient

Issue - State: closed - Opened by godisboy over 7 years ago - 1 comment

#11 - How to compute the WGAN loss gradient

Issue - State: closed - Opened by godisboy over 7 years ago - 1 comment

#10 - Where to get LSUN dataset

Issue - State: closed - Opened by tengyuliu over 7 years ago - 1 comment

#9 - Missmatch between loss in paper and code

Issue - State: closed - Opened by sguada over 7 years ago - 10 comments

#8 - Make sure the generator training always gets a full batch of noise

Pull Request - State: closed - Opened by FeepingCreature over 7 years ago - 1 comment

#7 - Instead of manually counting with str(i), use unique labels for dcgan

Pull Request - State: closed - Opened by FeepingCreature over 7 years ago - 1 comment

#6 - When specifying --noBN, use _nobn discriminator network as well

Pull Request - State: closed - Opened by FeepingCreature over 7 years ago - 1 comment

#5 - Typo in paper?

Issue - State: closed - Opened by PatWie over 7 years ago - 4 comments

#4 - Questions on Loss scale, Hyperparameters

Issue - State: open - Opened by lifelongeek over 7 years ago - 4 comments

#3 - critic output nonlinearity, loss and `n_critic`

Issue - State: closed - Opened by pavanramkumar over 7 years ago - 2 comments

#2 - Interpretation of Discriminator Loss

Issue - State: closed - Opened by LukasMosser over 7 years ago - 34 comments

#1 - trivial doc update

Pull Request - State: closed - Opened by black-puppydog over 7 years ago - 1 comment

GitHub / martinarjovsky/WassersteinGAN issues and pull requests