提交 982ab584 编写于 作者: L LDOUBLEV

merge fepan

上级 8c173feb
...@@ -39,7 +39,8 @@ class DSConv(nn.Layer): ...@@ -39,7 +39,8 @@ class DSConv(nn.Layer):
stride=1, stride=1,
groups=None, groups=None,
if_act=True, if_act=True,
act="relu"): act="relu",
**kwargs):
super(DSConv, self).__init__() super(DSConv, self).__init__()
if groups == None: if groups == None:
groups = in_channels groups = in_channels
...@@ -263,7 +264,7 @@ class CAFPN(nn.Layer): ...@@ -263,7 +264,7 @@ class CAFPN(nn.Layer):
class FEPAN(nn.Layer): class FEPAN(nn.Layer):
def __init__(self, in_channels, out_channels, **kwargs): def __init__(self, in_channels, out_channels, mode='large', **kwargs):
super(FEPAN, self).__init__() super(FEPAN, self).__init__()
self.out_channels = out_channels self.out_channels = out_channels
weight_attr = paddle.nn.initializer.KaimingUniform() weight_attr = paddle.nn.initializer.KaimingUniform()
...@@ -274,6 +275,15 @@ class FEPAN(nn.Layer): ...@@ -274,6 +275,15 @@ class FEPAN(nn.Layer):
self.pan_head_conv = nn.LayerList() self.pan_head_conv = nn.LayerList()
self.pan_lat_conv = nn.LayerList() self.pan_lat_conv = nn.LayerList()
if mode.lower() == 'lite':
p_layer = DSConv
elif mode.lower() == 'large':
p_layer = nn.Conv2D
else:
raise ValueError(
"mode can only be one of ['lite', 'large'], but received {}".
format(mode))
for i in range(len(in_channels)): for i in range(len(in_channels)):
self.ins_conv.append( self.ins_conv.append(
nn.Conv2D( nn.Conv2D(
...@@ -284,7 +294,7 @@ class FEPAN(nn.Layer): ...@@ -284,7 +294,7 @@ class FEPAN(nn.Layer):
bias_attr=False)) bias_attr=False))
self.inp_conv.append( self.inp_conv.append(
nn.Conv2D( p_layer(
in_channels=self.out_channels, in_channels=self.out_channels,
out_channels=self.out_channels // 4, out_channels=self.out_channels // 4,
kernel_size=9, kernel_size=9,
...@@ -303,7 +313,7 @@ class FEPAN(nn.Layer): ...@@ -303,7 +313,7 @@ class FEPAN(nn.Layer):
weight_attr=ParamAttr(initializer=weight_attr), weight_attr=ParamAttr(initializer=weight_attr),
bias_attr=False)) bias_attr=False))
self.pan_lat_conv.append( self.pan_lat_conv.append(
nn.Conv2D( p_layer(
in_channels=self.out_channels // 4, in_channels=self.out_channels // 4,
out_channels=self.out_channels // 4, out_channels=self.out_channels // 4,
kernel_size=9, kernel_size=9,
...@@ -346,86 +356,3 @@ class FEPAN(nn.Layer): ...@@ -346,86 +356,3 @@ class FEPAN(nn.Layer):
fuse = paddle.concat([p5, p4, p3, p2], axis=1) fuse = paddle.concat([p5, p4, p3, p2], axis=1)
return fuse return fuse
class FEPANLite(nn.Layer):
def __init__(self, in_channels, out_channels, **kwargs):
super(FEPANLite, self).__init__()
self.out_channels = out_channels
weight_attr = paddle.nn.initializer.KaimingUniform()
self.ins_conv = nn.LayerList()
self.inp_conv = nn.LayerList()
# pan head
self.pan_head_conv = nn.LayerList()
self.pan_lat_conv = nn.LayerList()
for i in range(len(in_channels)):
self.ins_conv.append(
nn.Conv2D(
in_channels=in_channels[i],
out_channels=self.out_channels,
kernel_size=1,
weight_attr=ParamAttr(initializer=weight_attr),
bias_attr=False))
self.inp_conv.append(
DSConv(
in_channels=self.out_channels,
out_channels=self.out_channels // 4,
kernel_size=9,
padding=4))
if i > 0:
self.pan_head_conv.append(
nn.Conv2D(
in_channels=self.out_channels // 4,
out_channels=self.out_channels // 4,
kernel_size=3,
padding=1,
stride=2,
weight_attr=ParamAttr(initializer=weight_attr),
bias_attr=False))
self.pan_lat_conv.append(
DSConv(
in_channels=self.out_channels // 4,
out_channels=self.out_channels // 4,
kernel_size=9,
padding=4))
def forward(self, x):
c2, c3, c4, c5 = x
in5 = self.ins_conv[3](c5)
in4 = self.ins_conv[2](c4)
in3 = self.ins_conv[1](c3)
in2 = self.ins_conv[0](c2)
out4 = in4 + F.upsample(
in5, scale_factor=2, mode="nearest", align_mode=1) # 1/16
out3 = in3 + F.upsample(
out4, scale_factor=2, mode="nearest", align_mode=1) # 1/8
out2 = in2 + F.upsample(
out3, scale_factor=2, mode="nearest", align_mode=1) # 1/4
f5 = self.inp_conv[3](in5)
f4 = self.inp_conv[2](out4)
f3 = self.inp_conv[1](out3)
f2 = self.inp_conv[0](out2)
pan3 = f3 + self.pan_head_conv[0](f2)
pan4 = f4 + self.pan_head_conv[1](pan3)
pan5 = f5 + self.pan_head_conv[2](pan4)
p2 = self.pan_lat_conv[0](f2)
p3 = self.pan_lat_conv[1](pan3)
p4 = self.pan_lat_conv[2](pan4)
p5 = self.pan_lat_conv[3](pan5)
p5 = F.upsample(p5, scale_factor=8, mode="nearest", align_mode=1)
p4 = F.upsample(p4, scale_factor=4, mode="nearest", align_mode=1)
p3 = F.upsample(p3, scale_factor=2, mode="nearest", align_mode=1)
fuse = paddle.concat([p5, p4, p3, p2], axis=1)
return fuse
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册