train

joint-ssgan/dataset.lua

@@ -19,10 +19,8 @@ require 'xlua'
 require 'image'
 
 local dataset = torch.class('dataLoader')
-list_file = '/nfs.yoda/xiaolonw/gan_render/genNormalCodeReg_high/trainlist_rand.txt'
+list_file = '../ssgan_models/trainlist_rand.txt'
 path_dataset = '/scratch/xiaolonw/render_data/'
--- path_dataset = '/nfs.yoda/xiaolonw/gan_render/data/'
-
 
 local initcheck = argcheck{
    pack=true,

joint-ssgan/test.lua

@@ -55,9 +55,9 @@ end
 -- model_G = model.G:cuda()
 -- model_G1= model.G1:cuda()
 
-model = torch.load('/nfs.yoda/xiaolonw/torch_projects/models/train_3dnormal_joint4/save/joint_Style_GAN.net')
+model = torch.load('../ssgan_models/joint_Style_GAN.net')
 model_G = model.G
-model = torch.load('/nfs.yoda/xiaolonw/torch_projects/models/dcgan_normal_72/save2/Structure_GAN.net')
+model = torch.load('../ssgan_models/Structure_GAN.net')
 model_G1 = model.G
 model_G = model_G:cuda()
 model_G1 = model_G1:cuda()
@@ -88,7 +88,9 @@ model_upsample:float()
 -- Get examples to plot
 function getSamples(dataset, N, beg)
 
-  local resultpath = '/nfs/hn38/users/xiaolonw/dcgan/joint_all_results/' 
+  local resultpath = '../results/joint_all_results/' 
+  os.execute('mkdir -p ' .. resultpath)
+
   local N = N or 8
   local noise_inputs = torch.Tensor(N, opt.noiseDim[1], opt.noiseDim[2], opt.noiseDim[3])
   local noise_inputs2 = torch.Tensor(N, opt.noiseDim[1], opt.noiseDim[2], opt.noiseDim[3])

joint-ssgan/train.lua

@@ -47,8 +47,8 @@ opt = lapp[[
 ]]
 
 if opt.gpu < 0 or opt.gpu > 8 then opt.gpu = false end
-opt.network1 = '/nfs.yoda/xiaolonw/torch_projects/models/dcgan_normal_72/save2/Structure_GAN.net'
-opt.network3 = '/nfs.yoda/xiaolonw/torch_projects/models/train_3dnormal_joint4/save/joint_Style_GAN.net'
+opt.network1 = '../ssgan_models/Structure_GAN.net'
+opt.network3 = '../ssgan_models/joint_Style_GAN.net'
 opt.pause = 0
 print(opt)
 

structure-gan/dataset.lua

@@ -19,7 +19,7 @@ require 'xlua'
 require 'image'
 
 local dataset = torch.class('dataLoader')
-list_file = '/nfs.yoda/xiaolonw/gan_render/genNormalCodeReg_high/trainlist_rand.txt'
+list_file = '../ssgan_models/trainlist_rand.txt'
 path_dataset = '/scratch/xiaolonw/render_data/'
 -- path_dataset = '/nfs.yoda/xiaolonw/gan_render/data/'
 

structure-gan/test.lua

@@ -50,7 +50,7 @@ else
   torch.setdefaulttensortype('torch.FloatTensor')
 end
 
-model = torch.load('/nfs.yoda/xiaolonw/torch_projects/models/dcgan_normal_72/adversarial_12.net') 
+model = torch.load('../ssgan_models/Structure_GAN.net') 
 model_G = model.G
 model_G = model_G:cuda()
 
@@ -62,7 +62,8 @@ opt.div_num = 127.5
 -- Get examples to plot
 function getSamples(dataset, N, beg)
 
-  local resultpath = '/nfs/hn38/users/xiaolonw/dcgan/test_structure_gan/' 
+  local resultpath = '../results/test_structure_gan/' 
+  os.execute('mkdir -p ' .. resultpath)
   local N = N or 8
   local noise_inputs = torch.Tensor(N, opt.noiseDim[1], opt.noiseDim[2], opt.noiseDim[3])
 

style-gan-fcn/dataset.lua

@@ -19,7 +19,7 @@ require 'xlua'
 require 'image'
 
 local dataset = torch.class('dataLoader')
-list_file = '/nfs.yoda/xiaolonw/gan_render/genNormalCodeReg_high/trainlist_rand.txt'
+list_file = '../ssgan_models/trainlist_rand.txt'
 path_dataset = '/scratch/xiaolonw/render_data/'
 -- path_dataset = '/nfs.yoda/xiaolonw/gan_render/data/'
 

style-gan-fcn/test_fcn.lua

@@ -55,9 +55,11 @@ end
 
 local list_file = '/nfs/hn46/xiaolonw/cnncode/viewer/testLabels.txt'
 local path_dataset = '/nfs/hn46/xiaolonw/cnncode/viewer/croptest/'
-local resultpath = '/nfs/hn38/users/xiaolonw/dcgan/ssgan_fcn/'
+local resultpath = '../results/ssgan_fcn/' 
+os.execute('mkdir -p ' .. resultpath)
 
-model = torch.load('/nfs.yoda/xiaolonw/torch_projects/models/train_3dnormal_fcn_cls/FCN.net')
+
+model = torch.load('../ssgan_models/FCN.net')
 -- model = torch.load('/nfs.yoda/xiaolonw/torch_projects/models/train_3dnormal_fcn_cls/fcn_5.net')
 model_FCN = model.FCN
 model_FCN = model_FCN:cuda()

style-gan-fcn/test_gan.lua

@@ -47,7 +47,7 @@ else
   torch.setdefaulttensortype('torch.FloatTensor')
 end
 
-model = torch.load('/nfs.yoda/xiaolonw/torch_projects/models/train_3dnormal_joint4/save/joint_Style_GAN.net') 
+model = torch.load('../ssgan_models/joint_Style_GAN.net') 
 model_G = model.G
 model_G = model_G:cuda()
 
@@ -61,11 +61,11 @@ opt.labelSize = 32
 paths.dofile('donkey.lua')
 
 
-
 -- Get examples to plot
 function getSamples(dataset, N, beg)
 
-  local resultpath = '/nfs/hn38/users/xiaolonw/dcgan/test_style_gan2/' 
+  local resultpath = '../results/test_style_gan2/' 
+  os.execute('mkdir -p ' .. resultpath)
   local N = N or 8
   local noise_inputs = torch.Tensor(N, opt.noiseDim[1], opt.noiseDim[2], opt.noiseDim[3])
   local cond_inputs = torch.Tensor(N, opt.condDim[1], opt.condDim[2], opt.condDim[3])

style-gan-nofcn/dataset.lua

@@ -19,7 +19,7 @@ require 'xlua'
 require 'image'
 
 local dataset = torch.class('dataLoader')
-list_file = '/nfs.yoda/xiaolonw/gan_render/genNormalCodeReg_high/trainlist_rand.txt'
+list_file = '../ssgan_models/trainlist_rand.txt'
 path_dataset = '/scratch/xiaolonw/render_data/'
 -- path_dataset = '/nfs.yoda/xiaolonw/gan_render/data/'
 

style-gan-nofcn/test.lua

@@ -47,7 +47,7 @@ else
   torch.setdefaulttensortype('torch.FloatTensor')
 end
 
-model = torch.load('/nfs.yoda/xiaolonw/torch_projects/models/train_3dnormal_rgb/bactch60/Style_GAN_nofcn.net') 
+model = torch.load('../ssgan_models/Style_GAN_nofcn.net') 
 model_G = model.G
 model_G = model_G:cuda()
 
@@ -64,7 +64,9 @@ paths.dofile('donkey.lua')
 -- Get examples to plot
 function getSamples(dataset, N, beg)
 
-  local resultpath = '/nfs/hn38/users/xiaolonw/dcgan/test_style_gan/' 
+
+  local resultpath = '../results/test_style_gan/' 
+  os.execute('mkdir -p ' .. resultpath)
   local N = N or 8
   local noise_inputs = torch.Tensor(N, opt.noiseDim[1], opt.noiseDim[2], opt.noiseDim[3])
   local cond_inputs = torch.Tensor(N, opt.condDim[1], opt.condDim[2], opt.condDim[3])