Implement ResGenv2
Implements a ResGenv2 architecture which slightly increases the complexity of the final output layer but causes it to be shared across all skip outputs.
This commit is contained in:
parent
1596a98493
commit
f217216c81
|
@ -78,19 +78,20 @@ class FixupResNet(nn.Module):
|
|||
# convs which are intended to repair artifacts caused by 2x interpolation.
|
||||
# This core layer should by itself accomplish 2x super-resolution. We use it in repeat to do the
|
||||
# requested SR.
|
||||
nf2 = int(num_filters/4)
|
||||
self.nf2 = int(num_filters/4)
|
||||
# This part isn't repeated. It de-filters the output from the previous step to fit the filter size used in the
|
||||
# upsampler-conv.
|
||||
self.upsampler_conv = nn.Conv2d(num_filters, nf2, kernel_size=3, stride=1, padding=1, bias=False)
|
||||
self.upsampler_conv = nn.Conv2d(num_filters, self.nf2, kernel_size=3, stride=1, padding=1, bias=False)
|
||||
self.uc_bias = nn.Parameter(torch.zeros(1))
|
||||
self.inplanes = nf2
|
||||
self.inplanes = self.nf2
|
||||
|
||||
# This is the repeated part.
|
||||
self.layer2 = self._make_layer(block, int(nf2), layers[1], stride=1, conv_type=conv5x5)
|
||||
self.skip2 = nn.Conv2d(nf2, 3, kernel_size=5, stride=1, padding=2, bias=False)
|
||||
self.skip2_bias = nn.Parameter(torch.zeros(1))
|
||||
if layers[1] > 0:
|
||||
# This is the repeated part.
|
||||
self.layer2 = self._make_layer(block, int(self.nf2), layers[1], stride=1, conv_type=conv5x5)
|
||||
self.skip2 = nn.Conv2d(self.nf2, 3, kernel_size=5, stride=1, padding=2, bias=False)
|
||||
self.skip2_bias = nn.Parameter(torch.zeros(1))
|
||||
|
||||
self.final_defilter = nn.Conv2d(nf2, 3, kernel_size=5, stride=1, padding=2, bias=True)
|
||||
self.final_defilter = nn.Conv2d(self.nf2, 3, kernel_size=5, stride=1, padding=2, bias=True)
|
||||
self.bias2 = nn.Parameter(torch.zeros(1))
|
||||
|
||||
for m in self.modules():
|
||||
|
@ -124,9 +125,12 @@ class FixupResNet(nn.Module):
|
|||
skip_lo = self.skip1(x) + self.skip1_bias
|
||||
|
||||
x = self.lrelu(self.upsampler_conv(x) + self.uc_bias)
|
||||
x = F.interpolate(x, scale_factor=2.0, mode='nearest')
|
||||
x = self.layer2(x)
|
||||
skip_med = self.skip2(x) + self.skip2_bias
|
||||
if self.upscale_applications > 0:
|
||||
x = F.interpolate(x, scale_factor=2.0, mode='nearest')
|
||||
x = self.layer2(x)
|
||||
skip_med = self.skip2(x) + self.skip2_bias
|
||||
else:
|
||||
skip_med = skip_lo
|
||||
|
||||
if self.upscale_applications > 1:
|
||||
x = F.interpolate(x, scale_factor=2.0, mode='nearest')
|
||||
|
@ -135,11 +139,59 @@ class FixupResNet(nn.Module):
|
|||
x = self.final_defilter(x) + self.bias2
|
||||
return x, skip_med, skip_lo
|
||||
|
||||
class FixupResNetV2(FixupResNet):
|
||||
def __init__(self, **kwargs):
|
||||
super(FixupResNetV2, self).__init__(**kwargs)
|
||||
# Use one unified filter-to-image stack, not the previous skip stacks.
|
||||
self.skip1 = None
|
||||
self.skip1_bias = None
|
||||
self.skip2 = None
|
||||
self.skip2_bias = None
|
||||
# The new filter-to-image stack will be 2 conv layers deep, not 1.
|
||||
self.final_process = nn.Conv2d(self.nf2, self.nf2, kernel_size=5, stride=1, padding=2, bias=True)
|
||||
self.bias2 = nn.Parameter(torch.zeros(1))
|
||||
self.fp_bn = nn.BatchNorm2d(self.nf2)
|
||||
self.final_defilter = nn.Conv2d(self.nf2, 3, kernel_size=3, stride=1, padding=1, bias=True)
|
||||
self.bias3 = nn.Parameter(torch.zeros(1))
|
||||
|
||||
def filter_to_image(self, filter):
|
||||
x = self.final_process(filter) + self.bias2
|
||||
x = self.lrelu(self.fp_bn(x))
|
||||
x = self.final_defilter(x) + self.bias3
|
||||
return x
|
||||
|
||||
def forward(self, x):
|
||||
x = self.conv1(x)
|
||||
x = self.lrelu(x + self.bias1)
|
||||
x = self.layer1(x)
|
||||
x = self.lrelu(self.upsampler_conv(x) + self.uc_bias)
|
||||
|
||||
skip_lo = self.filter_to_image(x)
|
||||
if self.upscale_applications > 0:
|
||||
x = F.interpolate(x, scale_factor=2.0, mode='nearest')
|
||||
x = self.layer2(x)
|
||||
skip_med = self.filter_to_image(x)
|
||||
|
||||
if self.upscale_applications > 1:
|
||||
x = F.interpolate(x, scale_factor=2.0, mode='nearest')
|
||||
x = self.layer2(x)
|
||||
|
||||
x = self.filter_to_image(x)
|
||||
return x, skip_med, skip_lo
|
||||
|
||||
def fixup_resnet34(nb_denoiser=20, nb_upsampler=10, **kwargs):
|
||||
"""Constructs a Fixup-ResNet-34 model.
|
||||
"""
|
||||
model = FixupResNet(FixupBasicBlock, [nb_denoiser, nb_upsampler], **kwargs)
|
||||
return model
|
||||
|
||||
def fixup_resnet34_v2(nb_denoiser=20, nb_upsampler=10, **kwargs):
|
||||
"""Constructs a Fixup-ResNet-34 model.
|
||||
"""
|
||||
kwargs['block'] = FixupBasicBlock
|
||||
kwargs['layers'] = [nb_denoiser, nb_upsampler]
|
||||
model = FixupResNetV2(**kwargs)
|
||||
return model
|
||||
|
||||
__all__ = ['FixupResNet', 'fixup_resnet34']
|
||||
|
||||
__all__ = ['FixupResNet', 'fixup_resnet34', 'fixup_resnet34_v2']
|
|
@ -36,6 +36,9 @@ def define_G(opt):
|
|||
elif which_model == 'ResGen':
|
||||
netG = ResGen_arch.fixup_resnet34(nb_denoiser=opt_net['nb_denoiser'], nb_upsampler=opt_net['nb_upsampler'],
|
||||
upscale_applications=opt_net['upscale_applications'], num_filters=opt_net['nf'])
|
||||
elif which_model == 'ResGenV2':
|
||||
netG = ResGen_arch.fixup_resnet34_v2(nb_denoiser=opt_net['nb_denoiser'], nb_upsampler=opt_net['nb_upsampler'],
|
||||
upscale_applications=opt_net['upscale_applications'], num_filters=opt_net['nf'])
|
||||
|
||||
# image corruption
|
||||
elif which_model == 'HighToLowResNet':
|
||||
|
|
Loading…
Reference in New Issue
Block a user