-
Notifications
You must be signed in to change notification settings - Fork 5
/
base_net_snow.py
304 lines (247 loc) · 10.4 KB
/
base_net_snow.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
from audioop import bias
from mimetypes import init
from pickle import TRUE
from tokenize import group
import torch
import torch.nn as nn
from einops import rearrange
import numbers
from timm.models.layers import DropPath,to_2tuple,trunc_normal_
import math
import torch.nn.functional as F
from condconv import *
def to_3d(x):
return rearrange(x, 'b c h w -> b (h w) c')
def to_4d(x,h,w):
return rearrange(x, 'b (h w) c -> b c h w',h=h,w=w)
class BiasFree_LayerNorm(nn.Module):
def __init__(self, normalized_shape):
super(BiasFree_LayerNorm, self).__init__()
if isinstance(normalized_shape, numbers.Integral):
normalized_shape = (normalized_shape,)
normalized_shape = torch.Size(normalized_shape)
assert len(normalized_shape) == 1
self.weight = nn.Parameter(torch.ones(normalized_shape))
self.normalized_shape = normalized_shape
def forward(self, x):
sigma = x.var(-1, keepdim=True, unbiased=False)
return x / torch.sqrt(sigma+1e-5) * self.weight
class WithBias_LayerNorm(nn.Module):
def __init__(self, normalized_shape):
super(WithBias_LayerNorm, self).__init__()
if isinstance(normalized_shape, numbers.Integral):
normalized_shape = (normalized_shape,)
normalized_shape = torch.Size(normalized_shape)
assert len(normalized_shape) == 1
self.weight = nn.Parameter(torch.ones(normalized_shape))
self.bias = nn.Parameter(torch.zeros(normalized_shape))
self.normalized_shape = normalized_shape
def forward(self, x):
mu = x.mean(-1, keepdim=True)
sigma = x.var(-1, keepdim=True, unbiased=False)
return (x - mu) / torch.sqrt(sigma+1e-5) * self.weight + self.bias
class LayerNorm(nn.Module):
def __init__(self, dim, LayerNorm_type):
super(LayerNorm, self).__init__()
if LayerNorm_type =='BiasFree':
self.body = BiasFree_LayerNorm(dim)
else:
self.body = WithBias_LayerNorm(dim)
def forward(self, x):
h, w = x.shape[-2:]
return to_4d(self.body(to_3d(x)), h, w)
class Down(nn.Module):
def __init__(self,in_channels,out_channels,kernel_size,stride,padding):
super(Down,self).__init__()
self.down = nn.Sequential(
nn.Conv2d(in_channels,out_channels,kernel_size=kernel_size,stride=stride,padding=padding,bias=False),
)
def forward(self,x):
x = self.down(x)
return x
class Up(nn.Module):
def __init__(self,in_channels,out_channels,kernel_size,stride,padding):
super(Up,self).__init__()
self.up = nn.Sequential(
nn.Sequential(nn.ConvTranspose2d(in_channels,out_channels,kernel_size=kernel_size,stride=stride,padding=padding))
)
def forward(self,x):
x = self.up(x)
return x
class UpSample(nn.Module):
def __init__(self, in_channels,out_channels,s_factor):
super(UpSample, self).__init__()
self.up = nn.Sequential(nn.Upsample(scale_factor=s_factor, mode='bilinear', align_corners=False),
nn.Conv2d(in_channels, out_channels, 1, stride=1, padding=0, bias=False))
def forward(self, x):
x = self.up(x)
return x
class DWconv(nn.Module):
def __init__(self,in_channels,out_channels):
super(DWconv,self).__init__()
self.dwconv = nn.Conv2d(in_channels,out_channels,3,1,1,bias=False,groups=in_channels)
def forward(self,x):
x = self.dwconv(x)
return x
class ChannelAttention(nn.Module):
def __init__(self,chns,factor,dynamic=False):
super().__init__()
self.avg_pool = nn.AdaptiveAvgPool2d(1)
if dynamic == False:
self.channel_map = nn.Sequential(
nn.Conv2d(chns,chns//factor,1,1,0),
nn.LeakyReLU(),
nn.Conv2d(chns//factor,chns,1,1,0),
nn.Sigmoid()
)
else:
self.channel_map = nn.Sequential(
CondConv2D(chns,chns//factor,1,1,0),
nn.LeakyReLU(),
CondConv2D(chns//factor,chns,1,1,0),
nn.Sigmoid()
)
def forward(self,x):
avg_pool = self.avg_pool(x)
map = self.channel_map(avg_pool)
return x*map
class LKA(nn.Module):
def __init__(self, dim):
super().__init__()
self.conv0 = nn.Conv2d(dim, dim, 5, padding=2, groups=dim)
self.act1 = nn.GELU()
self.conv_spatial = nn.Conv2d(dim, dim, 7, stride=1, padding=9, groups=dim, dilation=3)
self.act2 = nn.GELU()
self.conv1 = nn.Conv2d(dim, dim, 1)
def forward(self, x):
u = x.clone()
attn = self.conv0(x)
attn = self.act1(attn)
attn = self.conv_spatial(attn)
attn = self.act2(attn)
attn = self.conv1(attn)
return u * attn
class LKA_dynamic(nn.Module):
def __init__(self, dim):
super().__init__()
self.conv0 = CondConv2D(dim,dim,5,1,2,1,dim)
self.act1 = nn.GELU()
self.conv_spatial = CondConv2D(dim,dim,7,1,9,3,dim)
self.act2 = nn.GELU()
self.conv1 = nn.Conv2d(dim, dim, 1)
def forward(self, x):
u = x.clone()
attn = self.conv0(x)
attn = self.act1(attn)
attn = self.conv_spatial(attn)
attn = self.act2(attn)
attn = self.conv1(attn)
return u * attn
class Attention(nn.Module):
def __init__(self, d_model,dynamic=True):
super().__init__()
self.conv11 = nn.Conv2d(d_model,d_model,kernel_size=3,stride=1,padding=1,groups=d_model)
#self.activation = nn.GELU()
if dynamic == True:
self.spatial_gating_unit = LKA_dynamic(d_model)
else:
self.spatial_gating_unit = LKA(d_model)
self.conv22 = nn.Conv2d(d_model,d_model,kernel_size=3,stride=1,padding=1,groups=d_model)
def forward(self, x):
x = self.conv11(x)
x = self.spatial_gating_unit(x)
x = self.conv22(x)
return x
class ConvBlock(nn.Module):
def __init__(self, inp, oup, stride, expand_ratio,VAN=False,dynamic=False):
super(ConvBlock, self).__init__()
self.VAN = VAN
hidden_dim = round(inp * expand_ratio)
self.identity = stride == 1 and inp == oup
self.apply(self._init_weight)
if self.VAN == True:
if expand_ratio == 1:
self.conv = nn.Sequential(
LayerNorm(hidden_dim, 'BiasFree'),
Attention(hidden_dim,dynamic=dynamic),
)
else:
self.conv = nn.Sequential(
nn.Conv2d(inp, hidden_dim, 1, 1, 0 ),
LayerNorm(hidden_dim, 'BiasFree'),
Attention(hidden_dim,dynamic=dynamic),
nn.Conv2d(hidden_dim, oup, 1, 1, 0),
)
else:
if dynamic == False:
if expand_ratio == 1:
self.conv = nn.Sequential(
nn.Conv2d(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
LayerNorm(hidden_dim, 'BiasFree'),
nn.GELU(),
nn.Conv2d(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
ChannelAttention(hidden_dim,4,dynamic=dynamic),
nn.Conv2d(hidden_dim, oup, 1, 1, 0),
)
else:
self.conv = nn.Sequential(
# pw
nn.Conv2d(inp, hidden_dim, 1, 1, 0 ),
nn.Conv2d(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
LayerNorm(hidden_dim, 'BiasFree'),
nn.GELU(),
nn.Conv2d(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
ChannelAttention(hidden_dim,4,dynamic=dynamic),
nn.Conv2d(hidden_dim, oup, 1, 1, 0),
)
else:
if expand_ratio == 1:
self.conv = nn.Sequential(
CondConv2D(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
LayerNorm(hidden_dim, 'BiasFree'),
nn.GELU(),
CondConv2D(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
ChannelAttention(hidden_dim,4,dynamic=dynamic),
# pw-linear
nn.Conv2d(hidden_dim, oup, 1, 1, 0),
)
else:
self.conv = nn.Sequential(
nn.Conv2d(inp, hidden_dim, 1, 1, 0 ),
CondConv2D(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
LayerNorm(hidden_dim, 'BiasFree'),
nn.GELU(),
CondConv2D(hidden_dim, hidden_dim, 3, stride, 1,dilation=1, groups=hidden_dim),
ChannelAttention(hidden_dim,4,dynamic=dynamic),
nn.Conv2d(hidden_dim, oup, 1, 1, 0),
)
def _init_weight(self,m):
if isinstance(m,nn.Linear):
trunc_normal_(m.weight,std=0.02)
if isinstance(m,nn.Linear) and m.bias is not None:
nn.init.constant_(m.bias,0)
elif isinstance(m,nn.LayerNorm):
nn.init.constant_(m.bias,0)
nn.init.constant_(m.weight,1.0)
elif isinstance(m,nn.Conv2d):
fan_out = m.kernel_size[0]*m.kernel_size[1]*m.out_channels
fan_out //= m.groups
m.weight.data.normal_(0,math.sqrt(2.0/fan_out))
if m.bias is not None:
m.bias.data.zero_()
def forward(self, x):
if self.identity:
return x + self.conv(x)
else:
return self.conv(x)
class Conv_block(nn.Module):
def __init__(self,n,in_channel,out_channele,expand_ratio,VAN=False,dynamic=False):
super(Conv_block,self).__init__()
layers=[]
for i in range(n):
layers.append(ConvBlock(in_channel,out_channele,1 if i==0 else 1,expand_ratio,VAN=VAN,dynamic=dynamic))
in_channel = out_channele
self.model = nn.Sequential(*layers)
def forward(self,x):
x = self.model(x)
return x