-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathmodel_orig.py
176 lines (142 loc) · 6.86 KB
/
model_orig.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
import torch
import torch.nn as nn
import numpy as np
import torchvision
import torch.nn.functional as F
cfg = [64, 128, 256, 512, 512]
class Deep3d(nn.Module):
def __init__(self, in_channels=3, out_channels=3, device=torch.device('cpu')):
super(Deep3d, self).__init__()
self.device = device
vgg16 = torchvision.models.vgg16(pretrained=True)
modules = []
layer = []
for l in vgg16.features:
if isinstance(l, nn.MaxPool2d):
layer.append(l)
modules.append(layer)
layer = []
else:
layer.append(l)
scale = 1
deconv = []
layer = []
for m in range(len(modules)):
layer.append(nn.Conv2d(cfg[m], cfg[m], kernel_size=3, stride=1, padding=True))
layer.append(nn.ReLU(inplace=True))
layer.append(nn.Conv2d(cfg[m], cfg[m], kernel_size=3, stride=1, padding=True))
layer.append(nn.ReLU(inplace=True))
if(m==0):
layer.append(nn.ConvTranspose2d(cfg[m], 65, kernel_size=1, stride=1, padding=(0,0)))
# elif(m==len(modules)-1):
# scale *=4
# layer.append(nn.ConvTranspose2d(cfg[m], 65, kernel_size=scale*2, stride=scale, padding=(scale//2, scale//2)))
else:
scale *=2
layer.append(nn.ConvTranspose2d(cfg[m], 65, kernel_size=scale*2, stride=scale, padding=(scale//2, scale//2)))
deconv.append(layer)
layer = []
self.module_1 = nn.Sequential(*modules[0])
self.module_2 = nn.Sequential(*modules[1])
self.module_3 = nn.Sequential(*modules[2])
self.module_4 = nn.Sequential(*modules[3])
self.module_5 = nn.Sequential(*modules[4])
self.deconv_1 = nn.Sequential(*deconv[0])
self.deconv_2 = nn.Sequential(*deconv[1])
self.deconv_3 = nn.Sequential(*deconv[2])
self.deconv_4 = nn.Sequential(*deconv[3])
self.deconv_5 = nn.Sequential(*deconv[4])
self.linear_module = nn.Sequential(*[nn.Linear(15360,4096), # hyperparam choice
nn.ReLU(inplace=True),
nn.Dropout(p=0.5),
nn.Linear(4096,1950)]) # 1950=65(disparity range)*10*3(10*3 is feature map size)
# scale*=2
self.deconv_6 = nn.Sequential(*[nn.ConvTranspose2d(65,65,kernel_size=scale*2,stride=scale,padding=(scale//2,scale//2))])
self.upconv_final = nn.Sequential(*[nn.ConvTranspose2d(65,65,kernel_size=(4,4),stride=2,padding=(1,1)),
nn.ReLU(inplace=True),
nn.Conv2d(65,65,kernel_size=(3,3),stride=1,padding=(1,1)),
nn.Softmax(dim=1)])
for block in [self.deconv_1,self.deconv_2,self.deconv_3,self.deconv_4,self.deconv_5,self.deconv_6,self.linear_module,self.upconv_final]:
# for block in [self.deconv_1,self.deconv_2,self.deconv_3,self.deconv_4,self.deconv_5,self.deconv_6,self.upconv_final]:
for m in block:
if isinstance(m, nn.Conv2d):
nn.init.kaiming_normal_(m.weight, mode='fan_out', nonlinearity='relu')
if m.bias is not None:
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.BatchNorm2d):
nn.init.constant_(m.weight, 1)
nn.init.constant_(m.bias, 0)
elif isinstance(m, nn.Linear):
nn.init.normal_(m.weight, 0, 0.01)
nn.init.constant_(m.bias, 0)
def forward(self, x):
x_copy = x.clone()
pred = []
out_1 = self.module_1(x)
out_2 = self.module_2(out_1)
out_3 = self.module_3(out_2)
out_4 = self.module_4(out_3)
out_5 = self.module_5(out_4)
# print(out_5.shape)
out_5_flatten = out_5.view(x_copy.shape[0],-1)
out_6 = self.linear_module(out_5_flatten)
p1 = self.deconv_1(out_1)
p2 = self.deconv_2(out_2)
p3 = self.deconv_3(out_3)
p4 = self.deconv_4(out_4)
p5 = self.deconv_5(out_5)
# print(p5.shape)
p6 = self.deconv_6(out_6.view(x_copy.shape[0],65,3,10))
pred.append(p1)
pred.append(p2)
pred.append(p3)
pred.append(p4)
pred.append(p5)
pred.append(p6)
out = torch.zeros(pred[0].shape).to(self.device)
for p in pred:
# print('p',p.shape)
# print('out',out.shape)
out = torch.add(out, p)
out = self.upconv_final(out) # to be elt wise multiplied with shifted left views
new_right_image = torch.zeros(x_copy.size()).to(self.device)
stacked_shifted_view = None
stacked_out = None
for depth_map_idx in range(-33,32):
shifted_input_view = torch.zeros(x_copy.size()).to(self.device)
# print(shifted_input_view.size())
# print('x',x_copy.size())
if depth_map_idx<0:
shifted_input_view[:,:,:,:depth_map_idx] = x_copy[:,:,:,-depth_map_idx:]
elif depth_map_idx==0:
shifted_input_view = x_copy
else:
# print(depth_map_idx)
shifted_input_view[:,:,:,depth_map_idx:] = x_copy[:,:,:,:-depth_map_idx]
# print(shifted_input_view.shape)
if stacked_shifted_view is None:
stacked_shifted_view = shifted_input_view.unsqueeze(1)
else:
stacked_shifted_view = torch.cat((stacked_shifted_view,shifted_input_view.unsqueeze(1)),dim=1)
if stacked_out is None:
stacked_out = out[:,depth_map_idx+33:depth_map_idx+34,:,:].unsqueeze(1)
else:
stacked_out = torch.cat((stacked_out,out[:,depth_map_idx+33:depth_map_idx+34,:,:].unsqueeze(1)),dim=1)
# print("SO",stacked_out.shape)
# print(stacked_shifted_view.shape)
# new_right_image += torch.mul(shifted_input_view,out[:,depth_map_idx+33:depth_map_idx+34,:,:])
# return new_right_image
softmaxed_stacked_shifted_view = stacked_shifted_view
# softmaxed_stacked_shifted_view = F.softmax(stacked_shifted_view,dim=1)
# print(softmaxed_stacked_shifted_view.shape)
mult_soft_shift_out = torch.mul(stacked_out,softmaxed_stacked_shifted_view)
# print('mult',mult_soft_shift_out.shape)
final_rt_image = torch.sum(mult_soft_shift_out,dim=1)
# print(final_rt_image.shape)
return final_rt_image
if(__name__=='__main__'):
vgg16 = torchvision.models.vgg16(pretrained=True)
print(vgg16)
model = Deep3d().to(torch.device('cpu'))
out = model(torch.randn(10,3,96,320))
# model(torch.randn(1,3,320,96))