mrq/DL-Art-School
1
3
Fork 5
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Wrap vgg disc

Browse Source
This commit is contained in:
James Betker 2020-08-25 18:14:45 -06:00
parent 83f2f8d239
commit 434ed70a9a
1 changed files with 8 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

13
codes/models/networks.py
View File

@ -1,4 +1,5 @@
import torch import torch
import logging
import models.archs.SRResNet_arch as SRResNet_arch import models.archs.SRResNet_arch as SRResNet_arch
import models.archs.discriminator_vgg_arch as SRGAN_arch import models.archs.discriminator_vgg_arch as SRGAN_arch
import models.archs.DiscriminatorResnet_arch as DiscriminatorResnet_arch import models.archs.DiscriminatorResnet_arch as DiscriminatorResnet_arch
@ -16,6 +17,8 @@ import models.archs.arch_util as arch_util
import functools import functools
from collections import OrderedDict from collections import OrderedDict
logger = logging.getLogger('base')
# Generator # Generator
def define_G(opt, net_key='network_G', scale=None): def define_G(opt, net_key='network_G', scale=None):
if net_key is not None: if net_key is not None:
@ -143,11 +146,11 @@ def define_G(opt, net_key='network_G', scale=None):
class GradDiscWrapper(torch.nn.Module): class GradDiscWrapper(torch.nn.Module):
def __init__(self, m): def __init__(self, m):
super(GradDiscWrapper, self).__init__() super(GradDiscWrapper, self).__init__()
print("Wrapping a discriminator..") logger.info("Wrapping a discriminator..")
self.m = m self.m = m
def forward(self, x, lr): def forward(self, x):
return self.m(x, lr) return self.m(x)
def define_D_net(opt_net, img_sz=None, wrap=False): def define_D_net(opt_net, img_sz=None, wrap=False):
which_model = opt_net['which_model_D'] which_model = opt_net['which_model_D']
@ -156,6 +159,8 @@ def define_D_net(opt_net, img_sz=None, wrap=False):
netD = SRGAN_arch.Discriminator_VGG_128(in_nc=opt_net['in_nc'], nf=opt_net['nf'], input_img_factor=img_sz // 128, extra_conv=opt_net['extra_conv']) netD = SRGAN_arch.Discriminator_VGG_128(in_nc=opt_net['in_nc'], nf=opt_net['nf'], input_img_factor=img_sz // 128, extra_conv=opt_net['extra_conv'])
elif which_model == 'discriminator_vgg_128_gn': elif which_model == 'discriminator_vgg_128_gn':
netD = SRGAN_arch.Discriminator_VGG_128_GN(in_nc=opt_net['in_nc'], nf=opt_net['nf'], input_img_factor=img_sz // 128) netD = SRGAN_arch.Discriminator_VGG_128_GN(in_nc=opt_net['in_nc'], nf=opt_net['nf'], input_img_factor=img_sz // 128)
if wrap:
netD = GradDiscWrapper(netD)
elif which_model == 'discriminator_resnet': elif which_model == 'discriminator_resnet':
netD = DiscriminatorResnet_arch.fixup_resnet34(num_filters=opt_net['nf'], num_classes=1, input_img_size=img_sz) netD = DiscriminatorResnet_arch.fixup_resnet34(num_filters=opt_net['nf'], num_classes=1, input_img_size=img_sz)
elif which_model == 'discriminator_resnet_passthrough': elif which_model == 'discriminator_resnet_passthrough':
@ -173,8 +178,6 @@ def define_D_net(opt_net, img_sz=None, wrap=False):
final_temperature_step=opt_net['final_temperature_step']) final_temperature_step=opt_net['final_temperature_step'])
elif which_model == "cross_compare_vgg128": elif which_model == "cross_compare_vgg128":
netD = SRGAN_arch.CrossCompareDiscriminator(in_nc=opt_net['in_nc'], ref_channels=opt_net['ref_channels'], nf=opt_net['nf'], scale=opt_net['scale']) netD = SRGAN_arch.CrossCompareDiscriminator(in_nc=opt_net['in_nc'], ref_channels=opt_net['ref_channels'], nf=opt_net['nf'], scale=opt_net['scale'])
if wrap:
netD = GradDiscWrapper(netD)
else: else:
raise NotImplementedError('Discriminator model [{:s}] not recognized'.format(which_model)) raise NotImplementedError('Discriminator model [{:s}] not recognized'.format(which_model))
return netD return netD