实验笔记之——基于SRResNet的denoise
本博文是采用SRResNet来实现图像去噪的实验记录
python train.py -opt options/train/train_ne.json.json
setting（不知道为何，用我自己实现的srresnet需要两块GPU才可以跑起来，但是用xintao的可以一块就ok了，但都是跑得很慢~~~所以最终还是决定两块卡一起跑吧哎）
{
"name": "n_g_srresnet" //"001_RRDB_PSNR_x4_DIV2K" // please remove "debug_" during training or tensorboard wounld not work
,
"use_tb_logger": true,
"model": "sr",
//"crop_scale": 0,
"scale": 1//it must be 1
,
"gpu_ids": [4],
"datasets": {
"train": {
"name": "DIV2K800",
"mode": "LRHR" //it must be this, and the detail would be shown in LRHR_dataset.py
//, "noise_get": true///////////////////////////////////////////////////////////////////////
,
"dataroot_HR": "/home/guanwp/BasicSR_datasets/DIV2K800_sub" ///////////must be sub
,
"dataroot_LR": "/home/guanwp/BasicSR_datasets/DIV2K800_sub_n",
"subset_file": null,
"use_shuffle": true,
"n_workers": 8,
"batch_size": 16//32 //how many samples in each iters
,
"HR_size": 192 // 128 | 192
,
"use_flip": false //true//////////////////////////
,
"use_rot": false //true////////////////////////
},
"val": {
"name": "val_set5",
"mode": "LRHR",
"dataroot_HR": "/home/guanwp/BasicSR_datasets/val_set5/Set5",
"dataroot_LR": "/home/guanwp/BasicSR_datasets/val_set5/Set5_n"
//, "noise_get": true///////////////////////////////////////////////////////////////////////this is important
}
},
"path": {
"root": "/home/guanwp/BasicSR-master/",
"pretrain_model_G": null,
"experiments_root": "/home/guanwp/BasicSR-master/experiments/",
"models": "/home/guanwp/BasicSR-master/experiments/n_g_srresnet/models",
"log": "/home/guanwp/BasicSR-master/experiments/n_g_srresnet",
"val_images": "/home/guanwp/BasicSR-master/experiments/n_g_srresnet/val_images"
},
"network_G": {
"which_model_G": "srresnet"//"noise_estimation" //"espcn"//"srresnet"//"sr_resnet"//"fsrcnn"//"sr_resnet" // RRDB_net | sr_resnet
,
"norm_type": null,
"mode": "CNA",
"nf": 64 //56//64
,
"nb": 23,
"in_nc": 3,
"out_nc": 3,
"gc": 32,
"group": 1
},
"train": {
"lr_G": 1e-3//8e-4 //1e-3//2e-4
,
"lr_scheme": "MultiStepLR",
"lr_steps": [100000,200000,300000,400000,600000],
"lr_gamma": 0.5,
"pixel_criterion": "l2" //"l2_tv"//"l1"//'l2'//huber//Cross //should be MSE LOSS
,
"pixel_weight": 1.0,
"val_freq": 1e3,
"manual_seed": 0,
"niter": 8e5 //2e6//1e6
},
"logger": {
"print_freq": 200,
"save_checkpoint_freq": 1e3
}
}
网络结构
##########################################################################################################
#SRResNet, the Original version
#define the residual block
class O_Residual_Block(nn.Module):
def __init__(self):
super(O_Residual_Block,self).__init__()
self.conv1=nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, stride=1, padding=1, bias=False)
self.bn1=nn.BatchNorm2d(64, affine=True)
self.prelu=nn.PReLU()
self.conv2 = nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, stride=1, padding=1, bias=False)
self.bn2=nn.BatchNorm2d(64, affine=True)
def forward(self, x):
identity_data1 = x
output = self.prelu(self.bn1(self.conv1(x)))
output = self.bn2(self.conv2(output))
output = torch.add(output,identity_data1)
return output
##############################################
class OSRRESNET(nn.Module):
def __init__(self, in_nc, out_nc, nf, nb, upscale=2, norm_type='batch', act_type='relu', \
mode='NAC', res_scale=1, upsample_mode='upconv'):##play attention the upscales
super(OSRRESNET,self).__init__()
self.conv_input=nn.Conv2d(in_channels=in_nc,out_channels=nf,kernel_size=9,stride=1,padding=4,bias=False)
self.prelu=nn.PReLU()
self.residual=self.make_layer(O_Residual_Block,16)
self.conv_mid = nn.Conv2d(in_channels=nf, out_channels=nf, kernel_size=3, stride=1, padding=1, bias=False)
self.bn_mid=nn.BatchNorm2d(64, affine=True)
self.upscale4x = nn.Sequential(
nn.Conv2d(in_channels=64, out_channels=256, kernel_size=3, stride=1, padding=1, bias=False),
nn.PixelShuffle(2),
nn.PReLU(),
nn.Conv2d(in_channels=64, out_channels=256, kernel_size=3, stride=1, padding=1, bias=False),
nn.PixelShuffle(2),
nn.PReLU(),
)
self.conv_output = nn.Conv2d(in_channels=nf, out_channels=out_nc, kernel_size=9, stride=1, padding=4, bias=False)
def make_layer(self, block, num_of_layer):
layers = []
for _ in range(num_of_layer):
layers.append(block())
return nn.Sequential(*layers)
def forward(self,x):
out = self.prelu(self.conv_input(x))
residual = out
out = self.residual(out)
out = self.bn_mid(self.conv_mid(out))
out = torch.add(out,residual)
#out = self.upscale4x(out)
out = self.conv_output(out)
return out
##########################################################################################################################################################
然后会发现网络学到一半突然跑飞。。。。。。。。。。
网络改为
##########################################################################################################
#SRResNet, the Original version
#define the residual block
class DN_Residual_Block(nn.Module):
def __init__(self):
super(DN_Residual_Block,self).__init__()
self.conv1=nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, stride=1, padding=1, bias=False)
self.bn1=nn.BatchNorm2d(64, affine=True)
self.prelu=nn.PReLU()
self.conv2 = nn.Conv2d(in_channels=64, out_channels=64, kernel_size=3, stride=1, padding=1, bias=False)
self.bn2=nn.BatchNorm2d(64, affine=True)
def forward(self, x):
identity_data1 = x
output = self.prelu(self.bn1(self.conv1(x)))
output = self.bn2(self.conv2(output))
output = torch.add(output,identity_data1)
return output
##############################################
class DN_ResNet(nn.Module):
def __init__(self, in_nc, out_nc, nf, nb, upscale=2, norm_type='batch', act_type='relu', \
mode='NAC', res_scale=1, upsample_mode='upconv'):##play attention the upscales
super(DN_ResNet,self).__init__()
self.conv_input=nn.Conv2d(in_channels=in_nc,out_channels=nf,kernel_size=9,stride=1,padding=4,bias=False)
self.prelu=nn.PReLU()
self.residual=self.make_layer(O_Residual_Block,16)
self.conv_mid = nn.Conv2d(in_channels=nf, out_channels=nf, kernel_size=3, stride=1, padding=1, bias=False)
self.bn_mid=nn.BatchNorm2d(64, affine=True)
self.upscale4x = nn.Sequential(
nn.Conv2d(in_channels=64, out_channels=256, kernel_size=3, stride=1, padding=1, bias=False),
nn.PixelShuffle(2),
nn.PReLU(),
nn.Conv2d(in_channels=64, out_channels=256, kernel_size=3, stride=1, padding=1, bias=False),
nn.PixelShuffle(2),
nn.PReLU(),
)
self.conv_output = nn.Conv2d(in_channels=nf, out_channels=out_nc, kernel_size=9, stride=1, padding=4, bias=False)
def make_layer(self, block, num_of_layer):
layers = []
for _ in range(num_of_layer):
layers.append(block())
return nn.Sequential(*layers)
def forward(self,x):
out = self.prelu(self.conv_input(x))
residual = out
out = self.residual(out)
out = self.bn_mid(self.conv_mid(out))
out = torch.add(out,residual)
#out = self.upscale4x(out)
out = self.conv_output(out)
return out
##########################################################################################################################################################
settig改为：
{
"name": "srresnet" //"001_RRDB_PSNR_x4_DIV2K" // please remove "debug_" during training or tensorboard wounld not work
,
"use_tb_logger": true,
"model": "sr",
//"crop_scale": 0,
"scale": 1//it must be 1
,
"gpu_ids": [4,5],
"datasets": {
"train": {
"name": "DIV2K800",
"mode": "LRHR" //it must be this, and the detail would be shown in LRHR_dataset.py
//, "noise_get": true///////////////////////////////////////////////////////////////////////
,
"dataroot_HR": "/home/guanwp/BasicSR_datasets/DIV2K800_sub" ///////////must be sub
,
"dataroot_LR": "/home/guanwp/BasicSR_datasets/DIV2K800_sub_n",
"subset_file": null,
"use_shuffle": true,
"n_workers": 8,
"batch_size": 16//32 //how many samples in each iters
,
"HR_size": 128 // 128 | 192
,
"use_flip": false //true//////////////////////////
,
"use_rot": false //true////////////////////////
},
"val": {
"name": "val_set5",
"mode": "LRHR",
"dataroot_HR": "/home/guanwp/BasicSR_datasets/val_set5/Set5",
"dataroot_LR": "/home/guanwp/BasicSR_datasets/val_set5/Set5_n"
//, "noise_get": true///////////////////////////////////////////////////////////////////////this is important
}
},
"path": {
"root": "/home/guanwp/BasicSR-master/",
"pretrain_model_G": null,
"experiments_root": "/home/guanwp/BasicSR-master/experiments/",
"models": "/home/guanwp/BasicSR-master/experiments/srresnet/models",
"log": "/home/guanwp/BasicSR-master/experiments/srresnet",
"val_images": "/home/guanwp/BasicSR-master/experiments/srresnet/val_images"
},
"network_G": {
"which_model_G": "dn_srresnet"//"noise_estimation" //"espcn"//"srresnet"//"sr_resnet"//"fsrcnn"//"sr_resnet" // RRDB_net | sr_resnet
,
"norm_type": null,
"mode": "CNA",
"nf": 64 //56//64
,
"nb": 16,//////////////number of residual block
"in_nc": 3,
"out_nc": 3,
"gc": 32,
"group": 1
},
"train": {
"lr_G": 6e-4//8e-4 //1e-3//2e-4
,
"lr_scheme": "MultiStepLR",
"lr_steps": [200000,300000,400000,600000],
"lr_gamma": 0.5,
"pixel_criterion": "l1" //"l2_tv"//"l1"//'l2'//huber//Cross //should be MSE LOSS
,
"pixel_weight": 1.0,
"val_freq": 1e3,
"manual_seed": 0,
"niter": 8e5 //2e6//1e6
},
"logger": {
"print_freq": 200,
"save_checkpoint_freq": 1e3
}
}