File size: 7,449 Bytes
404d2af
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8b973ee
404d2af
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8b973ee
 
 
404d2af
 
 
 
 
8b973ee
 
 
404d2af
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8b973ee
404d2af
 
 
 
8b973ee
404d2af
 
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
"""
Hourglass network, taken from https://github.com/zhou13/lcnn
"""
import torch
import torch.nn as nn
import torch.nn.functional as F

__all__ = ["HourglassNet", "hg"]


class MultitaskHead(nn.Module):
    def __init__(self, input_channels, num_class):
        super(MultitaskHead, self).__init__()

        m = int(input_channels / 4)
        head_size = [[2], [1], [2]]
        heads = []
        for output_channels in sum(head_size, []):
            heads.append(
                nn.Sequential(
                    nn.Conv2d(input_channels, m, kernel_size=3, padding=1),
                    nn.ReLU(inplace=True),
                    nn.Conv2d(m, output_channels, kernel_size=1),
                )
            )
        self.heads = nn.ModuleList(heads)
        assert num_class == sum(sum(head_size, []))

    def forward(self, x):
        return torch.cat([head(x) for head in self.heads], dim=1)


class Bottleneck2D(nn.Module):
    expansion = 2

    def __init__(self, inplanes, planes, stride=1, downsample=None):
        super(Bottleneck2D, self).__init__()

        self.bn1 = nn.BatchNorm2d(inplanes)
        self.conv1 = nn.Conv2d(inplanes, planes, kernel_size=1)
        self.bn2 = nn.BatchNorm2d(planes)
        self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=stride, padding=1)
        self.bn3 = nn.BatchNorm2d(planes)
        self.conv3 = nn.Conv2d(planes, planes * 2, kernel_size=1)
        self.relu = nn.ReLU(inplace=True)
        self.downsample = downsample
        self.stride = stride

    def forward(self, x):
        residual = x

        out = self.bn1(x)
        out = self.relu(out)
        out = self.conv1(out)

        out = self.bn2(out)
        out = self.relu(out)
        out = self.conv2(out)

        out = self.bn3(out)
        out = self.relu(out)
        out = self.conv3(out)

        if self.downsample is not None:
            residual = self.downsample(x)

        out += residual

        return out


class Hourglass(nn.Module):
    def __init__(self, block, num_blocks, planes, depth):
        super(Hourglass, self).__init__()
        self.depth = depth
        self.block = block
        self.hg = self._make_hour_glass(block, num_blocks, planes, depth)

    def _make_residual(self, block, num_blocks, planes):
        layers = []
        for i in range(0, num_blocks):
            layers.append(block(planes * block.expansion, planes))
        return nn.Sequential(*layers)

    def _make_hour_glass(self, block, num_blocks, planes, depth):
        hg = []
        for i in range(depth):
            res = []
            for j in range(3):
                res.append(self._make_residual(block, num_blocks, planes))
            if i == 0:
                res.append(self._make_residual(block, num_blocks, planes))
            hg.append(nn.ModuleList(res))
        return nn.ModuleList(hg)

    def _hour_glass_forward(self, n, x):
        up1 = self.hg[n - 1][0](x)
        low1 = F.max_pool2d(x, 2, stride=2)
        low1 = self.hg[n - 1][1](low1)

        if n > 1:
            low2 = self._hour_glass_forward(n - 1, low1)
        else:
            low2 = self.hg[n - 1][3](low1)
        low3 = self.hg[n - 1][2](low2)
        # up2 = F.interpolate(low3, scale_factor=2)
        up2 = F.interpolate(low3, size=up1.shape[2:])
        out = up1 + up2
        return out

    def forward(self, x):
        return self._hour_glass_forward(self.depth, x)


class HourglassNet(nn.Module):
    """Hourglass model from Newell et al ECCV 2016"""

    def __init__(
        self, block, head, depth, num_stacks, num_blocks, num_classes, input_channels
    ):
        super(HourglassNet, self).__init__()

        self.inplanes = 64
        self.num_feats = 128
        self.num_stacks = num_stacks
        self.conv1 = nn.Conv2d(
            input_channels, self.inplanes, kernel_size=7, stride=2, padding=3
        )
        self.bn1 = nn.BatchNorm2d(self.inplanes)
        self.relu = nn.ReLU(inplace=True)
        self.layer1 = self._make_residual(block, self.inplanes, 1)
        self.layer2 = self._make_residual(block, self.inplanes, 1)
        self.layer3 = self._make_residual(block, self.num_feats, 1)
        self.maxpool = nn.MaxPool2d(2, stride=2)

        # build hourglass modules
        ch = self.num_feats * block.expansion
        # vpts = []
        hg, res, fc, score, fc_, score_ = [], [], [], [], [], []
        for i in range(num_stacks):
            hg.append(Hourglass(block, num_blocks, self.num_feats, depth))
            res.append(self._make_residual(block, self.num_feats, num_blocks))
            fc.append(self._make_fc(ch, ch))
            score.append(head(ch, num_classes))
            # vpts.append(VptsHead(ch))
            # vpts.append(nn.Linear(ch, 9))
            # score.append(nn.Conv2d(ch, num_classes, kernel_size=1))
            # score[i].bias.data[0] += 4.6
            # score[i].bias.data[2] += 4.6
            if i < num_stacks - 1:
                fc_.append(nn.Conv2d(ch, ch, kernel_size=1))
                score_.append(nn.Conv2d(num_classes, ch, kernel_size=1))
        self.hg = nn.ModuleList(hg)
        self.res = nn.ModuleList(res)
        self.fc = nn.ModuleList(fc)
        self.score = nn.ModuleList(score)
        # self.vpts = nn.ModuleList(vpts)
        self.fc_ = nn.ModuleList(fc_)
        self.score_ = nn.ModuleList(score_)

    def _make_residual(self, block, planes, blocks, stride=1):
        downsample = None
        if stride != 1 or self.inplanes != planes * block.expansion:
            downsample = nn.Sequential(
                nn.Conv2d(
                    self.inplanes,
                    planes * block.expansion,
                    kernel_size=1,
                    stride=stride,
                )
            )

        layers = []
        layers.append(block(self.inplanes, planes, stride, downsample))
        self.inplanes = planes * block.expansion
        for i in range(1, blocks):
            layers.append(block(self.inplanes, planes))

        return nn.Sequential(*layers)

    def _make_fc(self, inplanes, outplanes):
        bn = nn.BatchNorm2d(inplanes)
        conv = nn.Conv2d(inplanes, outplanes, kernel_size=1)
        return nn.Sequential(conv, bn, self.relu)

    def forward(self, x):
        out = []
        # out_vps = []
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)

        x = self.layer1(x)
        x = self.maxpool(x)
        x = self.layer2(x)
        x = self.layer3(x)

        for i in range(self.num_stacks):
            y = self.hg[i](x)
            y = self.res[i](y)
            y = self.fc[i](y)
            score = self.score[i](y)
            # pre_vpts = F.adaptive_avg_pool2d(x, (1, 1))
            # pre_vpts = pre_vpts.reshape(-1, 256)
            # vpts = self.vpts[i](x)
            out.append(score)
            # out_vps.append(vpts)
            if i < self.num_stacks - 1:
                fc_ = self.fc_[i](y)
                score_ = self.score_[i](score)
                x = x + fc_ + score_

        return out[::-1], y  # , out_vps[::-1]


def hg(**kwargs):
    model = HourglassNet(
        Bottleneck2D,
        head=kwargs.get("head", lambda c_in, c_out: nn.Conv2D(c_in, c_out, 1)),
        depth=kwargs["depth"],
        num_stacks=kwargs["num_stacks"],
        num_blocks=kwargs["num_blocks"],
        num_classes=kwargs["num_classes"],
        input_channels=kwargs["input_channels"],
    )
    return model