add InstanceNorm

.gitignore

@@ -141,6 +141,7 @@ test*/
 video_tmp/
 result/
 #./
+/pix2pix
 /pix2pixHD
 /tmp
 /to_make_show

models/pix2pix_model.py

@@ -97,7 +97,7 @@ def init_func(m):  # define the initialization function
             init.normal_(m.weight.data, 1.0, init_gain)
             init.constant_(m.bias.data, 0.0)
 
-    print('initialize network with %s' % init_type)
+    #print('initialize network with %s' % init_type)
     net.apply(init_func)  # apply the initialization function <init_func>
 
 

models/unet_model.py

@@ -31,4 +31,4 @@ def forward(self, x):
         x = self.up3(x, x2)
         x = self.up4(x, x1)
         x = self.outc(x)
-        return torch.sigmoid(x)
+        return torch.Tanh(x)

models/video_model.py

@@ -4,13 +4,23 @@
 from .unet_parts import *
 from .pix2pix_model import *
 
+Norm = 'batch'
+if Norm == 'instance':
+    NormLayer_2d = nn.InstanceNorm2d
+    NormLayer_3d = nn.InstanceNorm3d
+    use_bias = False
+else:
+    NormLayer_2d = nn.BatchNorm2d
+    NormLayer_3d = nn.BatchNorm3d
+    use_bias = True
+
 class encoder_2d(nn.Module):
     """Resnet-based generator that consists of Resnet blocks between a few downsampling/upsampling operations.
 
     We adapt Torch code and idea from Justin Johnson's neural style transfer project(https://github.com/jcjohnson/fast-neural-style)
     """
 
-    def __init__(self, input_nc, output_nc, ngf=64, norm_layer=nn.BatchNorm2d, use_dropout=False, n_blocks=6, padding_type='reflect'):
+    def __init__(self, input_nc, output_nc, ngf=64, norm_layer=NormLayer_2d, use_dropout=False, n_blocks=6, padding_type='reflect'):
         """Construct a Resnet-based generator
 
         Parameters:
@@ -55,7 +65,7 @@ class decoder_2d(nn.Module):
     We adapt Torch code and idea from Justin Johnson's neural style transfer project(https://github.com/jcjohnson/fast-neural-style)
     """
 
-    def __init__(self, input_nc, output_nc, ngf=64, norm_layer=nn.BatchNorm2d, use_dropout=False, n_blocks=6, padding_type='reflect'):
+    def __init__(self, input_nc, output_nc, ngf=64, norm_layer=NormLayer_2d, use_dropout=False, n_blocks=6, padding_type='reflect'):
         """Construct a Resnet-based generator
 
         Parameters:
@@ -114,8 +124,8 @@ class conv_3d(nn.Module):
     def __init__(self,inchannel,outchannel,kernel_size=3,stride=2,padding=1):
         super(conv_3d, self).__init__()
         self.conv = nn.Sequential(
-            nn.Conv3d(inchannel, outchannel, kernel_size=kernel_size, stride=stride, padding=padding, bias=False),
-            nn.BatchNorm3d(outchannel),
+            nn.Conv3d(inchannel, outchannel, kernel_size=kernel_size, stride=stride, padding=padding, bias=use_bias),
+            NormLayer_3d(outchannel),
             nn.ReLU(inplace=True),
         )
 
@@ -128,8 +138,8 @@ def __init__(self,inchannel,outchannel,kernel_size=3,stride=1,padding=1):
         super(conv_2d, self).__init__()
         self.conv = nn.Sequential(
             nn.ReflectionPad2d(padding),
-            nn.Conv2d(inchannel, outchannel, kernel_size=kernel_size, stride=stride, padding=0, bias=False),
-            nn.BatchNorm2d(outchannel),
+            nn.Conv2d(inchannel, outchannel, kernel_size=kernel_size, stride=stride, padding=0, bias=use_bias),
+            NormLayer_2d(outchannel),
             nn.ReLU(inplace=True),
         )
 
@@ -145,8 +155,8 @@ def __init__(self,in_channel):
         self.down2 = conv_3d(64, 128, 3, 2, 1)
         self.down3 = conv_3d(128, 256, 3, 1, 1)
         self.conver2d = nn.Sequential(
-            nn.Conv2d(256*int(in_channel/4), 256, kernel_size=3, stride=1, padding=1, bias=False),
-            nn.BatchNorm2d(256),
+            nn.Conv2d(256*int(in_channel/4), 256, kernel_size=3, stride=1, padding=1, bias=use_bias),
+            NormLayer_2d(256),
             nn.ReLU(inplace=True),
         )
 

train/add/train.py

@@ -50,8 +50,8 @@ def Toinputshape(imgs,masks,finesize):
         # print(imgs[i].shape,masks[i].shape)
         img,mask = data.random_transform_image(imgs[i], masks[i], finesize)
         # print(img.shape,mask.shape)
-        mask = mask.reshape(1,finesize,finesize)/255.0
-        img = img.transpose((2, 0, 1))/255.0
+        mask = (mask.reshape(1,finesize,finesize)/255.0-0.5)/0.5
+        img = (img.transpose((2, 0, 1))/255.0-0.5)/0.5
         result_imgs.append(img)
         result_masks.append(mask)
     result_imgs = np.array(result_imgs)

train/clean/train.py

@@ -18,21 +18,21 @@
 
 N = 25
 ITER = 10000000
-LR = 0.001
+LR = 0.0002
 beta1 = 0.5
 use_gpu = True
 use_gan = False
 use_L2 = True
-CONTINUE =  True
-lambda_L1 = 1.0#100.0
-lambda_gan = 1.0
+CONTINUE =  False
+lambda_L1 = 100.0
+lambda_gan = 1
 
 SAVE_FRE = 10000
 start_iter = 0
 finesize = 128
 loadsize = int(finesize*1.1)
-batchsize = 8
-perload_num = 32
+batchsize = 1
+perload_num = 16
 savename = 'MosaicNet_test'
 dir_checkpoint = 'checkpoints/'+savename
 util.makedirs(dir_checkpoint)
@@ -45,6 +45,7 @@
 videos = os.listdir('./dataset')
 videos.sort()
 lengths = []
+print('check dataset...')
 for video in videos:
     video_images = os.listdir('./dataset/'+video+'/ori')
     lengths.append(len(video_images))
@@ -55,7 +56,8 @@
 loadmodel.show_paramsnumber(netG,'netG')
 # netG = unet_model.UNet(3*N+1, 3)
 if use_gan:
-    netD = pix2pix_model.define_D(3*2+1, 64, 'basic', n_layers_D=3, norm='instance', init_type='normal', init_gain=0.02, gpu_ids=[])
+    #netD = pix2pix_model.define_D(3*2+1, 64, 'pixel', norm='instance')
+    netD = pix2pix_model.define_D(3*2+1, 64, 'basic', norm='instance')
     #netD = pix2pix_model.define_D(3*2+1, 64, 'n_layers', n_layers_D=5, norm='instance', init_type='normal', init_gain=0.02, gpu_ids=[])
 
 if CONTINUE:
@@ -104,26 +106,19 @@ def loaddata():
     return input_img,ground_true
 
 print('preloading data, please wait 5s...')
-# input_imgs=[]
-# ground_trues=[]
-input_imgs = torch.rand(batchsize,N*3+1,finesize,finesize).cuda()
-ground_trues = torch.rand(batchsize,3,finesize,finesize).cuda()
+
+if perload_num <= batchsize:
+    perload_num = batchsize*2
+input_imgs = torch.rand(perload_num,N*3+1,finesize,finesize).cuda()
+ground_trues = torch.rand(perload_num,3,finesize,finesize).cuda()
 load_cnt = 0
 
 def preload():
     global load_cnt   
     while 1:
         try:
-            # input_img,ground_true = loaddata()
-            # input_imgs.append(input_img)
-            # ground_trues.append(ground_true)
-            ran = random.randint(0, batchsize-1)
+            ran = random.randint(0, perload_num-1)
             input_imgs[ran],ground_trues[ran] = loaddata()
-
-
-            # if len(input_imgs)>perload_num:
-            #     del(input_imgs[0])
-            #     del(ground_trues[0])
             load_cnt += 1
             # time.sleep(0.1)
         except Exception as e:
@@ -133,21 +128,24 @@ def preload():
 t = threading.Thread(target=preload,args=())  #t为新创建的线程
 t.daemon = True
 t.start()
-while load_cnt < batchsize*2:
+
+time_start=time.time()
+while load_cnt < perload_num:
     time.sleep(0.1)
+time_end=time.time()
+print('load speed:',round((time_end-time_start)/perload_num,3),'s/it')
+
 
+util.copyfile('./train.py', os.path.join(dir_checkpoint,'train.py'))
+util.copyfile('../../models/video_model.py', os.path.join(dir_checkpoint,'model.py'))
 netG.train()
 time_start=time.time()
 print("Begin training...")
 for iter in range(start_iter+1,ITER):
 
-    # inputdata,target = loaddata()
-    # ran = random.randint(1, perload_num-2)
-    # inputdata = inputdatas[ran]
-    # target = targets[ran]
-
-    inputdata = input_imgs.clone()
-    target = ground_trues.clone()
+    ran = random.randint(0, perload_num-batchsize-1)
+    inputdata = input_imgs[ran:ran+batchsize].clone()
+    target = ground_trues[ran:ran+batchsize].clone()
 
     pred = netG(inputdata)
 
@@ -262,13 +260,13 @@ def preload():
         netG.eval()
 
         test_names = os.listdir('./test')
+        test_names.sort()
         result = np.zeros((finesize*2,finesize*len(test_names),3), dtype='uint8')
 
         for cnt,test_name in enumerate(test_names,0):
             img_names = os.listdir(os.path.join('./test',test_name,'image'))
             img_names.sort()
             inputdata = np.zeros((finesize,finesize,3*N+1), dtype='uint8')
-            img_names.sort()
             for i in range(0,N):
                 img = impro.imread(os.path.join('./test',test_name,'image',img_names[i]))
                 img = impro.resize(img,finesize)
@@ -286,4 +284,4 @@ def preload():
             result[finesize:finesize*2,finesize*cnt:finesize*(cnt+1),:] = pred
 
         cv2.imwrite(os.path.join(dir_checkpoint,str(iter+1)+'_test.png'), result)
-        netG.train()
+        netG.train()

util/data.py

@@ -74,11 +74,11 @@ def random_transform_video(src,target,finesize,N):
         target = target[:,::-1,:]
 
     #random color
-    alpha = random.uniform(-0.2,0.2)
+    alpha = random.uniform(-0.3,0.3)
     beta  = random.uniform(-0.2,0.2)
-    b     = random.uniform(-0.1,0.1)
-    g     = random.uniform(-0.1,0.1)
-    r     = random.uniform(-0.1,0.1)
+    b     = random.uniform(-0.05,0.05)
+    g     = random.uniform(-0.05,0.05)
+    r     = random.uniform(-0.05,0.05)
     for i in range(N):
         src[:,:,i*3:(i+1)*3] = color_adjust(src[:,:,i*3:(i+1)*3],alpha,beta,b,g,r)
     target = color_adjust(target,alpha,beta,b,g,r)

util/util.py

@@ -79,4 +79,10 @@ def get_bar(percent,num = 25):
         else:
             bar += '-'
     bar += ']'
-    return bar+' '+str(round(percent,2))+'%'
+    return bar+' '+str(round(percent,2))+'%'
+
+def copyfile(scr,dst):
+    try:
+        shutil.copyfile(src, dst)
+    except Exception as e:
+        print(e)