wtjiang98 / beautygan_pytorch Goto Github PK

Hi, thanks for your implement work, but i doubt wheater the implement of makeup loss is correct, the paper says that we should calculate the histogram matching between src image and ref image firstly, then calculate the l2 norm between fake image which generated by generator and matched image, i
find that you just calculate makeuo loss between the generated image and reference image which dont normlization the value to [-1, 1], and use L1 norm, could you help me? thanks!

Hello, I am interested in your work. I have cloned your code and downloaded the BeautyGAN dataset(and move it to the directory of your code). But when I run 'test.py' / 'train.py', it shows the error--No such file or directory: './data/images/train_SYMIX.txt'. I wonder what the .txt means, and I really hope that you can give some guide to run the code.

ps: the dataset is like: makeup_dataset-->all-->images丶segs-->makeup丶non-makeup. the pics in "segs" are all just black.

Hi, does anyone know why the BeautyGAN authors choose 6 residual blocks in the generator but not 9 as in CycleGAN because the training images are 256*256?

I also find that some unofficial implementations based on tensorflow are using 9 blocks. For example, https://github.com/baldFemale/beautyGAN-tf-Implement

It's really weird!

hi, is there a pretrained model for beautyGAN ?

Excuse , I have a question, what does the para --cls_list means ?

It seems that website of beautyGan has disappeared, could you please share the dataset?

您好！想请问一下MT-dataset数据集里面的segs里面的mask的分割代码开源了吗？

Could you upload the pre-trained vgg weight or let me know how to get the weight?

can you explain the process of face parsing?
the project and the pretrained model?

Could you share the vgg model, please?

guys，Looks like the dataset link is broken。who can put it on google drive or somewhere i can download，thanks very much

After many hours, I finally can run the code 2333. Here are some tips to run the code:

1. the train_MAKEMIX.txt means the names of those with-makeup pics in the MT dataset(train_SYMIX means non-makeup pics). It looks like this:
   

The names are duplicated in every row because the mask pics share the same name. The mask pics are in the "seg" folder of the MT dataset.

You can wirte a Python script to automatically read the names of the pics. As for me, I choose 2400 makeup pics for training, the rest 300 for testing. Remember to duplicate the name!

2. You can organize the dataset like this:
   
   Then you should change the paths in makeup.py. For example:
   

3. You can just download the VGG model from the Pytorch model zoo

import torchvision.models as models

#self.vgg = net.VGG()

#self.vgg.load_state_dict(torch.load('vgg_conv.pth')) 

self.vgg=models.vgg16(pretrained=True)

And then, write a forward function on your own to seize the 4th conv layer:

#you can print the vgg16 model and find that the 4th layer conv's id is 17.

def vgg_forward(self,model,x):

        for i in range(18):

            x=model.features[i](x)

        return x

Finally:

vgg_org=self.vgg_forward(self.vgg,org_A)

vgg_org = Variable(vgg_org.data).detach()

vgg_fake_A=self.vgg_forward(self.vgg,fake_A)

g_loss_A_vgg = self.criterionL2(vgg_fake_A, vgg_org) * self.lambda_A * self.lambda_vgg

......

(At this time the network speed of my home really sucks....It is so hard to download the ImageNet dataset, and it's hard to make the parameters match since the author have made some modifications to the VGG. I think the method above can work for you~)

At last, I am really grateful for the work that the author has done. It helps me a lot! Great thanks!

Hi, thanks for your share of your amazing work!

Do you have any plan to update the project code in November?

Hello，I'm very interested in this. But I ran the code, there are so many options. And I aslo download the database.But I am not sure which args means non-make up picture,please help