guim3 / icgan Goto Github PK

The resource disappeared. I can not find the Pre-trained CelebA model. Could you please re-upload a new one?

Just curious why you chose to implement in Lua, not by Python tensorflow or Pytorch?

I love your work and I want to replicate it for my research. Which file should I modify if I want different attributes from the one in the paper? For example, changing the skin complexion and also eye sizes

Discriminator and generator error have different error scales, so discriminator error can't be appreciated. Having two independent y-axis will make both error displays visible.

Can you tell where i have to change if i only want smiling face of any input image for example in attributeTransfer.lua file we pass two image and it swaps the attribute of that but what i input an single image and want only the 17th column smiling face of that.
Can you suggest me how to do that,

I get the error below on running th trainGAN.lua

/IcGAN/data/data.lua:32: attempt to compare number with nil

and the error of data.lua is in
9: function data.new(n, dataset_name, opt_)
32: if n > 0 then

I find that the n in the 'data.lua' equals nil, which is defined in trainGAN.lua by opt.nThreads
32: local data = DataLoader.new(opt.nThreads, opt.dataset, opt)

but i can't find the definition of opt.nThreads in code,then n equals nil.
Please suggest what could possibly be wrong.

Hi Guim, thank you for this enlightening model. I am trying to re-implement the mnist generation image in your paper, which fixed latent code Z and modify label Y to generate different hand-written numbers.

I can see that you haven't provided instructions in README, nor could I find any pre-trained models. So right now I am using 'reconstructionWithVariations.lua' with slight modifications to support mnist. Also, I am using the generator of epoch 25 and encoder of epoch 15 for both Z and Y, trained exactly as instructed by README.

So could your please tell me if you used the code from other files? And if not, which pre-trained models did you use? Thanks a lot!