def __init__(self,
inplanes,
planes,
stride=1,
downsample=None,
groups=1,
base_width=64,
dilation=1,
norm_layer=None):
super(BasicBlock, self).__init__()
if norm_layer is None:
norm_layer = nn.ComplexBatchNorm2d
if groups != 1 or base_width != 64:
raise ValueError(
'BasicBlock only supports groups=1 and base_width=64')
if dilation > 1:
raise NotImplementedError(
"Dilation > 1 not supported in BasicBlock")
# Both self.conv1 and self.downsample layers downsample the input when stride != 1
self.conv1 = conv3x3(inplanes, planes, stride)
self.bn1 = norm_layer(planes)
self.relu = nn.ComplexReLU()
self.conv2 = conv3x3(planes, planes)
self.bn2 = norm_layer(planes)
self.downsample = downsample
self.stride = stride
def __init__(self, in_channels, out_channels, kernel_size=3, upsampling=1):
conv2d = nn.ComplexConv2d(in_channels,
out_channels,
kernel_size=kernel_size,
padding=kernel_size // 2)
upsampling = nn.ComplexUpsamplingBilinear2d(
scale_factor=upsampling) if upsampling > 1 else nn.Identity()
activation = nn.ComplexReLU() # Always
super().__init__(conv2d, upsampling, activation)
def __init__(
self,
in_channels,
out_channels,
kernel_size,
padding=0,
stride=1,
use_batchnorm=True,
):
# if use_batchnorm == "inplace" and InPlaceABN is None:
# raise RuntimeError(
# "In order to use `use_batchnorm='inplace'` inplace_abn package must be installed. "
# + "To install see: https://github.com/mapillary/inplace_abn"
# )
conv = nn.ComplexConv2d(
in_channels,
out_channels,
kernel_size,
stride=stride,
padding=padding,
bias=not (use_batchnorm),
)
relu = nn.ComplexReLU()
if use_batchnorm == "inplace":
# bn = InPlaceABN(out_channels, activation="leaky_relu", activation_param=0.0)
# relu = nn.Identity()
pass
elif use_batchnorm and use_batchnorm != "inplace":
bn = nn.ComplexBatchNorm2d(out_channels)
else:
bn = nn.Identity()
super(Conv2dReLU, self).__init__(conv, bn, relu)
def __init__(self,
inplanes,
planes,
stride=1,
downsample=None,
groups=1,
base_width=64,
dilation=1,
norm_layer=None):
super(Bottleneck, self).__init__()
if norm_layer is None:
norm_layer = nn.ComplexBatchNorm2d
width = int(planes * (base_width / 64.)) * groups
# Both self.conv2 and self.downsample layers downsample the input when stride != 1
self.conv1 = conv1x1(inplanes, width)
self.bn1 = norm_layer(width)
self.conv2 = conv3x3(width, width, stride, groups, dilation)
self.bn2 = norm_layer(width)
self.conv3 = conv1x1(width, planes * self.expansion)
self.bn3 = norm_layer(planes * self.expansion)
self.relu = nn.ComplexReLU()
self.downsample = downsample
self.stride = stride
def __init__(self,
block,
layers,
num_classes=1000,
zero_init_residual=False,
groups=1,
width_per_group=64,
replace_stride_with_dilation=None,
norm_layer=None):
super(ResNet, self).__init__()
if norm_layer is None:
norm_layer = nn.ComplexBatchNorm2d
self._norm_layer = norm_layer
self.inplanes = 64
self.dilation = 1
if replace_stride_with_dilation is None:
# each element in the tuple indicates if we should replace
# the 2x2 stride with a dilated convolution instead
replace_stride_with_dilation = [False, False, False]
if len(replace_stride_with_dilation) != 3:
raise ValueError("replace_stride_with_dilation should be None "
"or a 3-element tuple, got {}".format(
replace_stride_with_dilation))
self.groups = groups
self.base_width = width_per_group
self.conv1 = nn.ComplexConv2d(3,
self.inplanes,
kernel_size=7,
stride=2,
padding=3,
bias=False)
self.bn1 = norm_layer(self.inplanes)
self.relu = nn.ComplexReLU()
self.maxpool = nn.ComplexMaxPool2d(kernel_size=3, stride=2, padding=1)
self.layer1 = self._make_layer(block, 64, layers[0])
self.layer2 = self._make_layer(block,
128,
layers[1],
stride=2,
dilate=replace_stride_with_dilation[0])
self.layer3 = self._make_layer(block,
256,
layers[2],
stride=2,
dilate=replace_stride_with_dilation[1])
self.layer4 = self._make_layer(block,
512,
layers[3],
stride=2,
dilate=replace_stride_with_dilation[2])
self.avgpool = nn.ComplexAdaptiveAvgPool2d((1, 1))
self.fc = nn.ComplexLinear(512 * block.expansion, num_classes)
#DEBUG: Can drop the following if it doesn't work
for m in self.modules():
if isinstance(m, nn.ComplexConv2d):
init.kaiming_normal_(m.conv_r.weight,
mode='fan_out',
nonlinearity='relu')
init.kaiming_normal_(m.conv_i.weight,
mode='fan_out',
nonlinearity='relu')
elif isinstance(
m, (nn.ComplexBatchNorm2d)
): # , nn.GroupNorm)): # ComplexGroupNorm not yet implemented
init.constant_(m.weight, 1)
init.constant_(m.bias, 0)
# Zero-initialize the last BN in each residual branch,
# so that the residual branch starts with zeros, and each residual block behaves like an identity.
# This improves the model by 0.2~0.3% according to https://arxiv.org/abs/1706.02677
if zero_init_residual:
for m in self.modules():
if isinstance(m, Bottleneck):
init.constant_(m.bn3.weight, 0)
elif isinstance(m, BasicBlock):
init.constant_(m.bn2.weight, 0)