-
Notifications
You must be signed in to change notification settings - Fork 21
/
Copy pathmodel.py
193 lines (160 loc) · 6.39 KB
/
model.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
import os
import sys
import random
import math
import time
import h5py
import cv2
import pickle
import torch
import numpy as np
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
import torch.utils.data as data
from torch.utils.data import DataLoader, Dataset
from torchvision import datasets, transforms
from utils.util import *
class myDatasets(Dataset):
def __init__(self, data_path):
self.crop_size = config.crop_size
self.channels = 3 * len(config.sample_idx)
self.landmark_size = config.landmark_num * 2
self.h5f = h5py.File(data_path, 'r')
self.feature = self.h5f['feature']
self.landmark = self.h5f['landmark']
self.metadata = self.h5f['metadata']
self.sample_len = self.metadata.shape[0]
self.zeros_mark = np.zeros((self.landmark_size,), dtype=np.float32)
def __len__(self):
return self.sample_len
def __getitem__(self, index):
feature = self.feature[index, :]
landmark = self.landmark[index, :]
metadata = self.metadata[index, :]
feat_data = feature.reshape((self.crop_size, self.crop_size, self.channels))
feat_data = feat_data.transpose(2, 0, 1)
face_label = metadata[1]
mark_label = self.zeros_mark if face_label == 0 else landmark[:self.landmark_size]
return feat_data, mark_label, face_label
class WingLoss(nn.Module):
def __init__(self, omega=10, epsilon=2):
super(WingLoss, self).__init__()
self.omega = omega
self.epsilon = epsilon
def forward(self, pred, target):
y = target
y_hat = pred
delta_y = (y - y_hat).abs()
delta_y1 = delta_y[delta_y < self.omega]
delta_y2 = delta_y[delta_y >= self.omega]
loss1 = self.omega * torch.log(1 + delta_y1 / self.epsilon)
C = self.omega - self.omega * math.log(1 + self.omega / self.epsilon)
loss2 = delta_y2 - C
return (loss1.sum() + loss2.sum()) / (len(loss1) + len(loss2))
def weights_init(m):
classname = m.__class__.__name__
if 'Conv2d' in classname:
nn.init.xavier_normal_(m.weight.data)
nn.init.constant_(m.bias.data, 0.0)
elif 'Linear' in classname:
nn.init.xavier_normal_(m.weight)
nn.init.constant_(m.bias, 0.0)
class PNET(nn.Module):
def __init__(self):
super(PNET, self).__init__()
groups = len(config.step1_idx)
in_channels = groups * 3
dims = [3, 16, 32, 64]
self.m1 = nn.Sequential(
nn.BatchNorm2d(in_channels),
nn.Conv2d(in_channels, groups * dims[1], kernel_size=3, stride=1, padding=0, groups=groups, bias=True),
nn.BatchNorm2d(groups * dims[1]),
nn.ReLU(),
nn.MaxPool2d(kernel_size=2)
)
self.m2 = nn.Sequential(
nn.Conv2d(groups * dims[1], groups * dims[2], kernel_size=3, stride=1, padding=1, groups=groups, bias=True),
nn.BatchNorm2d(groups * dims[2]),
nn.ReLU(),
nn.MaxPool2d(kernel_size=3, stride=2)
)
self.m3 = nn.Sequential(
nn.Conv2d(groups * dims[2], groups * dims[3], kernel_size=2, stride=1, padding=0, groups=groups, bias=True),
nn.ReLU()
)
self.pd = nn.Sequential(
nn.Linear(groups * dims[3], 256),
nn.ReLU(),
nn.Linear(256, 2 * config.landmark_num)
)
def forward(self, x):
x = self.m1(x)
x = self.m2(x)
x = self.m3(x)
x = x.view(x.size(0), -1)
x = self.pd(x)
return x
class RNET(nn.Module):
def __init__(self, is_train=True):
super(RNET, self).__init__()
groups = len(config.step2_idx)
in_channels = groups * 3
dims = [3, 8, 16, 32]
# dims = [3, 16, 32, 64]
self.m1 = nn.Sequential(
nn.BatchNorm2d(in_channels),
nn.Conv2d(in_channels, groups * dims[1], kernel_size=3, stride=1, padding=0, dilation=1, groups=groups, bias=True),
nn.BatchNorm2d(groups * dims[1]),
nn.ReLU(),
nn.MaxPool2d(kernel_size=2)
)
self.m2 = nn.Sequential(
nn.Conv2d(groups * dims[1], groups * dims[2], kernel_size=3, stride=1, padding=1, dilation=1, groups=groups, bias=True),
nn.BatchNorm2d(groups * dims[2]),
nn.ReLU(),
nn.MaxPool2d(kernel_size=2)
)
self.m3 = nn.Sequential(
nn.Conv2d(groups * dims[2], groups * dims[3], kernel_size=2, stride=1, padding=0, dilation=1, groups=groups, bias=True),
nn.ReLU()
)
self.m4 = nn.Sequential(
nn.Linear(groups * dims[3], 256),
nn.ReLU()
)
self.m5 = nn.Linear(256, 2 * config.landmark_num)
self.m6 = nn.Linear(256, 2)
self.m7 = nn.LogSoftmax(dim=1) if is_train else nn.Softmax(dim=1)
def forward(self, x):
x = self.m1(x)
x = self.m2(x)
x = self.m3(x)
x = x.view(x.size(0), -1)
x = self.m4(x)
y1 = self.m5(x)
x = self.m6(x)
y2 = self.m7(x)
return y1, y2
def get_pmodel(model_path=None):
is_cuda = not config.no_cuda and torch.cuda.is_available()
device = torch.device('cuda' if is_cuda else 'cpu')
model = PNET()
if model_path is not None:
checkpoint = torch.load(model_path, map_location='cpu')
model.load_state_dict(checkpoint)
else:
model.apply(weights_init)
model = model.to(device)
return model, device
def get_rmodel(model_path=None, is_train=True):
is_cuda = not config.no_cuda and torch.cuda.is_available()
device = torch.device('cuda' if is_cuda else 'cpu')
model = RNET(is_train=is_train)
if model_path is not None:
checkpoint = torch.load(model_path, map_location='cpu')
model.load_state_dict(checkpoint)
else:
model.apply(weights_init)
model = model.to(device)
return model, device