/
util.py
164 lines (130 loc) · 5.5 KB
/
util.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
import torch
import torch.nn as nn
import torch.nn.functional as F
import numpy as np
import sklearn
from sklearn.metrics import accuracy_score, f1_score
def AU_detection_eval_src(loader, base_net, au_enc, use_gpu=True):
missing_label = 999
for i, batch in enumerate(loader):
input, label = batch
if use_gpu:
input, label = input.cuda(), label.cuda()
base_feat = base_net(input)
au_feat, au_output = au_enc(base_feat)
au_output = F.sigmoid(au_output)
if i == 0:
all_output = au_output.data.cpu().float()
all_label = label.data.cpu().float()
else:
all_output = torch.cat((all_output, au_output.data.cpu().float()), 0)
all_label = torch.cat((all_label, label.data.cpu().float()), 0)
AUoccur_pred_prob = all_output.data.numpy()
AUoccur_actual = all_label.data.numpy()
AUoccur_pred = np.zeros(AUoccur_pred_prob.shape)
AUoccur_pred[AUoccur_pred_prob < 0.5] = 0
AUoccur_pred[AUoccur_pred_prob >= 0.5] = 1
AUoccur_actual = AUoccur_actual.transpose((1, 0))
AUoccur_pred = AUoccur_pred.transpose((1, 0))
f1score_arr = np.zeros(AUoccur_actual.shape[0])
acc_arr = np.zeros(AUoccur_actual.shape[0])
for i in range(AUoccur_actual.shape[0]):
curr_actual = AUoccur_actual[i]
curr_pred = AUoccur_pred[i]
new_curr_actual = curr_actual[curr_actual != missing_label]
new_curr_pred = curr_pred[curr_actual != missing_label]
f1score_arr[i] = f1_score(new_curr_actual, new_curr_pred)
acc_arr[i] = accuracy_score(new_curr_actual, new_curr_pred)
return f1score_arr, acc_arr
def AU_detection_eval_tgt(loader, base_net, land_enc, au_enc, invar_shape_enc, feat_gen, use_gpu=True):
missing_label = 999
for i, batch in enumerate(loader):
input, label = batch
if use_gpu:
input, label = input.cuda(), label.cuda()
base_feat = base_net(input)
align_attention, align_feat, align_output = land_enc(base_feat)
invar_shape_output = invar_shape_enc(base_feat)
new_gen = feat_gen(align_attention, invar_shape_output)
new_gen_au_feat, new_gen_au_output = au_enc(new_gen)
au_output = F.sigmoid(new_gen_au_output)
if i == 0:
all_output = au_output.data.cpu().float()
all_label = label.data.cpu().float()
else:
all_output = torch.cat((all_output, au_output.data.cpu().float()), 0)
all_label = torch.cat((all_label, label.data.cpu().float()), 0)
AUoccur_pred_prob = all_output.data.numpy()
AUoccur_actual = all_label.data.numpy()
AUoccur_pred = np.zeros(AUoccur_pred_prob.shape)
AUoccur_pred[AUoccur_pred_prob < 0.5] = 0
AUoccur_pred[AUoccur_pred_prob >= 0.5] = 1
AUoccur_actual = AUoccur_actual.transpose((1, 0))
AUoccur_pred = AUoccur_pred.transpose((1, 0))
f1score_arr = np.zeros(AUoccur_actual.shape[0])
acc_arr = np.zeros(AUoccur_actual.shape[0])
for i in range(AUoccur_actual.shape[0]):
curr_actual = AUoccur_actual[i]
curr_pred = AUoccur_pred[i]
new_curr_actual = curr_actual[curr_actual != missing_label]
new_curr_pred = curr_pred[curr_actual != missing_label]
f1score_arr[i] = f1_score(new_curr_actual, new_curr_pred)
acc_arr[i] = accuracy_score(new_curr_actual, new_curr_pred)
return f1score_arr, acc_arr
def land_softmax_loss(input, target, weight=None, size_average=True, reduce=True):
classify_loss = nn.CrossEntropyLoss(weight=weight, size_average=size_average, reduce=reduce)
for i in range(input.size(1)):
t_input = input[:, i, :, :]
t_input = t_input.view(t_input.size(0), -1)
t_target = target[:, i]
t_loss = classify_loss(t_input, t_target)
t_loss = torch.unsqueeze(t_loss, 0)
if i == 0:
loss = t_loss
else:
loss = torch.cat((loss, t_loss), 0)
if size_average:
return loss.mean()
else:
return loss.sum()
def land_adaptation_loss(input, size_average=True, reduce=True):
classify_loss = nn.MSELoss(size_average=size_average, reduce=reduce)
use_gpu = torch.cuda.is_available()
for i in range(input.size(1)):
t_input = input[:, i, :, :]
t_input = t_input.view(t_input.size(0), -1)
t_target = torch.ones(t_input.size()) * 1.0 / t_input.size(1)
if use_gpu:
t_target = t_target.cuda()
t_loss = classify_loss(t_input, t_target)
t_loss = torch.unsqueeze(t_loss, 0)
if i == 0:
loss = t_loss
else:
loss = torch.cat((loss, t_loss), 0)
if size_average:
return loss.mean()
else:
return loss.sum()
def land_discriminator_loss(input, target, size_average=True, reduce=True):
classify_loss = nn.MSELoss(size_average=size_average, reduce=reduce)
use_gpu = torch.cuda.is_available()
for i in range(input.size(1)):
t_input = input[:, i, :, :]
t_input = t_input.view(t_input.size(0), -1)
t_target = torch.zeros(t_input.size())
if use_gpu:
t_target = t_target.cuda()
t_true_target = target[:, i]
for j in range(t_true_target.size(0)):
t_target[j, t_true_target[j]] = 1
t_loss = classify_loss(t_input, t_target)
t_loss = torch.unsqueeze(t_loss, 0)
if i == 0:
loss = t_loss
else:
loss = torch.cat((loss, t_loss), 0)
if size_average:
return loss.mean()
else:
return loss.sum()