def __init__(self,
in_channels,
out_channels,
kernel_size,
stride=1,
groups=1,
is_vd_mode=False,
act=None,
name=None, ):
super(ConvBNLayer, self).__init__()
self.is_vd_mode = is_vd_mode
self.act = act
self._pool2d_avg = nn.AvgPool2d(
kernel_size=2, stride=2, padding=0, ceil_mode=True)
self._conv = nn.Conv2d(
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=(kernel_size - 1) // 2,
groups=groups,
bias=False)
if name == "conv1":
bn_name = "bn_" + name
else:
bn_name = "bn" + name[3:]
self._batch_norm = nn.BatchNorm2d(
out_channels,
track_running_stats=True,
)
if act is not None:
self._act = Activation(act_type=act, inplace=True)
def __init__(self,
in_channels,
out_channels,
kernel_size,
stride,
groups=1,
if_act=True,
act=None,
name=None):
super(ConvBNLayer, self).__init__()
self.if_act = if_act
self.conv = nn.Conv2d(
in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=(kernel_size - 1) // 2,
groups=groups,
bias=False)
self.bn = nn.BatchNorm2d(
out_channels,)
self.act = act
if act is not None:
self._act = Activation(act)
def __init__(self,
in_channels,
out_channels,
kernel_size=4,
stride=2,
padding=1,
groups=1,
if_act=True,
act=None,
name=None):
super(DeConvBNLayer, self).__init__()
self.if_act = if_act
self.act = act
self.deconv = nn.ConvTranspose2d(in_channels=in_channels,
out_channels=out_channels,
kernel_size=kernel_size,
stride=stride,
padding=padding,
groups=groups,
bias=False)
self.bn = nn.BatchNorm2d(out_channels)
self.act = act
if self.act is not None:
self._act = Activation(act_type=self.act, inplace=True)
def __init__(self, in_channels, reduction=4, name=""):
super(SEModule, self).__init__()
self.avg_pool = nn.AdaptiveAvgPool2d(1)
self.conv1 = nn.Conv2d(in_channels=in_channels,
out_channels=in_channels // reduction,
kernel_size=1,
stride=1,
padding=0,
bias=True)
self.relu1 = Activation(act_type='relu', inplace=True)
self.conv2 = nn.Conv2d(in_channels=in_channels // reduction,
out_channels=in_channels,
kernel_size=1,
stride=1,
padding=0,
bias=True)
self.hard_sigmoid = Activation(act_type='hard_sigmoid', inplace=True)
def __init__(self, in_channels, **kwargs):
super(Head, self).__init__()
self.conv1 = nn.Conv2d(in_channels=in_channels,
out_channels=in_channels // 4,
kernel_size=3,
padding=1,
bias=False)
self.conv_bn1 = nn.BatchNorm2d(in_channels // 4)
self.relu1 = Activation(act_type='relu')
self.conv2 = nn.ConvTranspose2d(in_channels=in_channels // 4,
out_channels=in_channels // 4,
kernel_size=2,
stride=2)
self.conv_bn2 = nn.BatchNorm2d(in_channels // 4)
self.relu2 = Activation(act_type='relu')
self.conv3 = nn.ConvTranspose2d(in_channels=in_channels // 4,
out_channels=1,
kernel_size=2,
stride=2)