/
helper_func.py
268 lines (191 loc) · 10.7 KB
/
helper_func.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
import torch
import torch.nn.functional as F
def val_gzsl(test_X, test_label, target_classes,in_package,bias = 0):
batch_size = in_package['batch_size']
model = in_package['model']
device = in_package['device']
with torch.no_grad():
start = 0
ntest = test_X.size()[0]
predicted_label = torch.LongTensor(test_label.size())
for i in range(0, ntest, batch_size):
end = min(ntest, start+batch_size)
input = test_X[start:end].to(device)
out_package = model(input)
output = out_package['S_pp']
output[:,target_classes] = output[:,target_classes]+bias
predicted_label[start:end] = torch.argmax(output.data, 1)
start = end
acc = compute_per_class_acc_gzsl(test_label, predicted_label, target_classes, in_package)
return acc
def map_label(label, classes):
mapped_label = torch.LongTensor(label.size()).fill_(-1)
for i in range(classes.size(0)):
mapped_label[label==classes[i]] = i
return mapped_label
def val_zs_gzsl(test_X, test_label, unseen_classes,in_package,bias = 0):
batch_size = in_package['batch_size']
model = in_package['model']
device = in_package['device']
with torch.no_grad():
start = 0
ntest = test_X.size()[0]
predicted_label_gzsl = torch.LongTensor(test_label.size())
predicted_label_zsl = torch.LongTensor(test_label.size())
predicted_label_zsl_t = torch.LongTensor(test_label.size())
for i in range(0, ntest, batch_size):
end = min(ntest, start+batch_size)
input = test_X[start:end].to(device)
out_package = model(input)
output = out_package['S_pp']
output_t = output.clone()
output_t[:,unseen_classes] = output_t[:,unseen_classes]+torch.max(output)+1
predicted_label_zsl[start:end] = torch.argmax(output_t.data, 1)
predicted_label_zsl_t[start:end] = torch.argmax(output.data[:,unseen_classes], 1)
output[:,unseen_classes] = output[:,unseen_classes]+bias
predicted_label_gzsl[start:end] = torch.argmax(output.data, 1)
start = end
acc_gzsl = compute_per_class_acc_gzsl(test_label, predicted_label_gzsl, unseen_classes, in_package)
acc_zs = compute_per_class_acc_gzsl(test_label, predicted_label_zsl, unseen_classes, in_package)
acc_zs_t = compute_per_class_acc(map_label(test_label, unseen_classes), predicted_label_zsl_t, unseen_classes.size(0))
return acc_gzsl,acc_zs_t
def compute_per_class_acc(test_label, predicted_label, nclass):
acc_per_class = torch.FloatTensor(nclass).fill_(0)
for i in range(nclass):
idx = (test_label == i)
acc_per_class[i] = torch.sum(test_label[idx]==predicted_label[idx]).float() / torch.sum(idx).float()
return acc_per_class.mean().item()
def compute_per_class_acc_gzsl(test_label, predicted_label, target_classes, in_package):
device = in_package['device']
per_class_accuracies = torch.zeros(target_classes.size()[0]).float().to(device).detach()
predicted_label = predicted_label.to(device)
for i in range(target_classes.size()[0]):
is_class = test_label == target_classes[i]
per_class_accuracies[i] = torch.div((predicted_label[is_class]==test_label[is_class]).sum().float(),is_class.sum().float())
return per_class_accuracies.mean().item()
def eval_zs_gzsl(dataloader,model,device,bias_seen=0, bias_unseen=0, batch_size=50):
model.eval()
# print('bias_seen {} bias_unseen {}'.format(bias_seen,bias_unseen))
test_seen_feature = dataloader.data['test_seen']['resnet_features']
test_seen_label = dataloader.data['test_seen']['labels'].to(device)
test_unseen_feature = dataloader.data['test_unseen']['resnet_features']
test_unseen_label = dataloader.data['test_unseen']['labels'].to(device)
seenclasses = dataloader.seenclasses
unseenclasses = dataloader.unseenclasses
batch_size = batch_size
in_package = {'model':model,'device':device, 'batch_size':batch_size}
with torch.no_grad():
acc_seen = val_gzsl(test_seen_feature, test_seen_label, seenclasses, in_package,bias=bias_seen)
acc_novel,acc_zs = val_zs_gzsl(test_unseen_feature, test_unseen_label, unseenclasses, in_package,bias = bias_unseen)
if (acc_seen+acc_novel)>0:
H = (2*acc_seen*acc_novel) / (acc_seen+acc_novel)
else:
H = 0
return acc_seen, acc_novel, H, acc_zs
def val_gzsl_k(k,test_X, test_label, target_classes,in_package,bias = 0,is_detect=False):
batch_size = in_package['batch_size']
model = in_package['model']
device = in_package['device']
n_classes = in_package["num_class"]
with torch.no_grad():
start = 0
ntest = test_X.size()[0]
test_label = F.one_hot(test_label, num_classes=n_classes)
predicted_label = torch.LongTensor(test_label.size()).fill_(0).to(test_label.device)
for i in range(0, ntest, batch_size):
end = min(ntest, start+batch_size)
input = test_X[start:end].to(device)
out_package = model(input)
output = out_package['S_pp']
output[:,target_classes] = output[:,target_classes]+bias
_,idx_k = torch.topk(output,k,dim=1)
if is_detect:
assert k == 1
detection_mask=in_package["detection_mask"]
predicted_label[start:end] = detection_mask[torch.argmax(output.data, 1)]
else:
predicted_label[start:end] = predicted_label[start:end].scatter_(1,idx_k,1)
start = end
acc = compute_per_class_acc_gzsl_k(test_label, predicted_label, target_classes, in_package)
return acc
def val_zs_gzsl_k(k,test_X, test_label, unseen_classes,in_package,bias = 0,is_detect=False):
batch_size = in_package['batch_size']
model = in_package['model']
device = in_package['device']
n_classes = in_package["num_class"]
with torch.no_grad():
start = 0
ntest = test_X.size()[0]
test_label_gzsl = F.one_hot(test_label, num_classes=n_classes)
predicted_label_gzsl = torch.LongTensor(test_label_gzsl.size()).fill_(0).to(test_label.device)
predicted_label_zsl = torch.LongTensor(test_label.size())
predicted_label_zsl_t = torch.LongTensor(test_label.size())
for i in range(0, ntest, batch_size):
end = min(ntest, start+batch_size)
input = test_X[start:end].to(device)
out_package = model(input)
output = out_package['S_pp']
output_t = output.clone()
output_t[:,unseen_classes] = output_t[:,unseen_classes]+torch.max(output)+1
predicted_label_zsl[start:end] = torch.argmax(output_t.data, 1)
predicted_label_zsl_t[start:end] = torch.argmax(output.data[:,unseen_classes], 1)
output[:,unseen_classes] = output[:,unseen_classes]+bias
_,idx_k = torch.topk(output,k,dim=1)
if is_detect:
assert k == 1
detection_mask=in_package["detection_mask"]
predicted_label_gzsl[start:end] = detection_mask[torch.argmax(output.data, 1)]
else:
predicted_label_gzsl[start:end] = predicted_label_gzsl[start:end].scatter_(1,idx_k,1)
start = end
acc_gzsl = compute_per_class_acc_gzsl_k(test_label_gzsl, predicted_label_gzsl, unseen_classes, in_package)
#print('acc_zs: {} acc_zs_t: {}'.format(acc_zs,acc_zs_t))
return acc_gzsl,-1
def compute_per_class_acc_k(test_label, predicted_label, nclass):
acc_per_class = torch.FloatTensor(nclass).fill_(0)
for i in range(nclass):
idx = (test_label == i)
acc_per_class[i] = torch.sum(test_label[idx]==predicted_label[idx]).float() / torch.sum(idx).float()
return acc_per_class.mean().item()
def compute_per_class_acc_gzsl_k(test_label, predicted_label, target_classes, in_package):
device = in_package['device']
per_class_accuracies = torch.zeros(target_classes.size()[0]).float().to(device).detach()
predicted_label = predicted_label.to(device)
hit = test_label*predicted_label
for i in range(target_classes.size()[0]):
target = target_classes[i]
n_pos = torch.sum(hit[:,target])
n_gt = torch.sum(test_label[:,target])
per_class_accuracies[i] = torch.div(n_pos.float(),n_gt.float())
#pdb.set_trace()
return per_class_accuracies.mean().item()
def eval_zs_gzsl_k(k,dataloader,model,device,bias_seen,bias_unseen,is_detect=False):
model.eval()
print('bias_seen {} bias_unseen {}'.format(bias_seen,bias_unseen))
test_seen_feature = dataloader.data['test_seen']['resnet_features']
test_seen_label = dataloader.data['test_seen']['labels'].to(device)
test_unseen_feature = dataloader.data['test_unseen']['resnet_features']
test_unseen_label = dataloader.data['test_unseen']['labels'].to(device)
seenclasses = dataloader.seenclasses
unseenclasses = dataloader.unseenclasses
batch_size = 100
n_classes = dataloader.ntrain_class+dataloader.ntest_class
in_package = {'model':model,'device':device, 'batch_size':batch_size,'num_class':n_classes}
if is_detect:
print("Measure novelty detection k: {}".format(k))
detection_mask = torch.zeros((n_classes,n_classes)).long().to(dataloader.device)
detect_label = torch.zeros(n_classes).long().to(dataloader.device)
detect_label[seenclasses]=1
detection_mask[seenclasses,:] = detect_label
detect_label = torch.zeros(n_classes).long().to(dataloader.device)
detect_label[unseenclasses]=1
detection_mask[unseenclasses,:]=detect_label
in_package["detection_mask"]=detection_mask
with torch.no_grad():
acc_seen = val_gzsl_k(k,test_seen_feature, test_seen_label, seenclasses, in_package,bias=bias_seen,is_detect=is_detect)
acc_novel,acc_zs = val_zs_gzsl_k(k,test_unseen_feature, test_unseen_label, unseenclasses, in_package,bias = bias_unseen,is_detect=is_detect)
if (acc_seen+acc_novel)>0:
H = (2*acc_seen*acc_novel) / (acc_seen+acc_novel)
else:
H = 0
return acc_seen, acc_novel, H, acc_zs