GAN ( ECCV 2016 )样式和结构

・3 分钟阅读

  • 源代码名称: ss-gan
  • 源代码网址: https://www.github.com/xiaolonw/ss-gan
  • ss-gan的文档
  • ss-gan的源代码下载
  • Git URL:
    git://www.github.com/xiaolonw/ss-gan.git
  • Git Clone代码到本地:
    git clone https://www.github.com/xiaolonw/ss-gan
  • Subversion代码到本地:
    $ svn co --depth empty https://www.github.com/xiaolonw/ss-gan
                              Checked out revision 1.
                              $ cd repo
                              $ svn up trunk
              
  • ss-gan

    此代码是基于Torch的eyescream项目开发的: 项目站点

    此代码是针对S^2-GAN的训练和测试实现,以下文章:

    Xiaolong和Abhinav Gupta ,基于风格和结构对抗性网络的生成图像建模,PROC欧洲计算机视觉会议,2016年,pdf

    BibTeX :

    
    @inproceedings{Wang_SSGAN2016,
    
    
     Author = {Xiaolong Wang and Abhinav Gupta},
    
    
     Title = {Generative Image Modeling using Style and Structure Adversarial Networks},
    
    
     Booktitle = {ECCV},
    
    
     Year = {2016},
    
    
    }
    
    
    
    

    模型和数据集

    这些训练的模型可以从使用代码的一般说明

    要进行训练,需要执行以下操作:

    
     Update the path_dataset = '/scratch/xiaolonw/render_data/' in dataset.lua 
    
    
     Update the opt.save in train.lua for saving models 
    
    
    
    

    为了进行测试,可以将模型下载到ssgan_models文件夹中。

    结构gan

    结构gan的代码在结构gan中:

    
     train.lua: training Stucture-GAN
    
    
     test.lua: testing Stucture-GAN
    
    
     ssgan_models/Structure_GAN.net is our trained model
    
    
    
    

    Style-GAN

    没有FCN约束的Style-GAN代码在style-gan-nofcn中:

    
     train.lua: training Style-GAN
    
    
     test.lua: testing Style-GAN (To run this you need to download the dataset)
    
    
     ssgan_models/Style_GAN_nofcn.net is our trained model
    
    
    
    

    具有FCN约束的Style-GAN的代码在style-gan-fcn中:

    
     train_fcn.lua: training FCN for surface normal estimation
    
    
     test_fcn.lua: testing FCN for surface normal estimation (To run this you need to download the dataset)
    
    
     ssgan_models/FCN.net is our trained model
    
    
    
     train_gan.lua: training Style-GAN
    
    
     test_gan.lua: testing Style-GAN (To run this you need to download the dataset)
    
    
     ssgan_models/joint_Style_GAN.net is our trained model
    
    
    
    

    S^2-GAN的联合学习

    
     train.lua: joint learning 
    
    
     test.lua: testing S^2-GAN
    
    
     ssgan_models/joint_SSGAN.net is our trained model
    
    
    
    
    讨论
    Fansisi profile image