-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_mnist-03_09-29-2020-17:41:58.log
101 lines (100 loc) · 4.7 KB
/
test_mnist-03_09-29-2020-17:41:58.log
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
[INFO 2020-09-29 17:41:58,917 root]
meta:
data_path: ~/mount/
gpu_id: '7'
model_name: mnist-03
save_path: ./
seed: 2019
test:
batch_size: 500
dataset: mnist
epsilon: 0.3
network: basic
num_test_samples: 10000
[INFO 2020-09-29 17:41:58,917 root] Preparing data...
[INFO 2020-09-29 17:42:03,969 root] Building model...
[INFO 2020-09-29 17:42:06,492 root] Starting attack...
[INFO 2020-09-29 18:03:39,219 root] AA acc: 0.9200.
500 successfully perturbed
apgd-ce - 2/20 - 37 out of 500 successfully perturbed
apgd-ce - 3/20 - 39 out of 500 successfully perturbed
apgd-ce - 4/20 - 47 out of 500 successfully perturbed
apgd-ce - 5/20 - 41 out of 500 successfully perturbed
apgd-ce - 6/20 - 44 out of 500 successfully perturbed
apgd-ce - 7/20 - 39 out of 500 successfully perturbed
apgd-ce - 8/20 - 51 out of 500 successfully perturbed
apgd-ce - 9/20 - 50 out of 500 successfully perturbed
apgd-ce - 10/20 - 27 out of 500 successfully perturbed
apgd-ce - 11/20 - 10 out of 500 successfully perturbed
apgd-ce - 12/20 - 34 out of 500 successfully perturbed
apgd-ce - 13/20 - 15 out of 500 successfully perturbed
apgd-ce - 14/20 - 21 out of 500 successfully perturbed
apgd-ce - 15/20 - 15 out of 500 successfully perturbed
apgd-ce - 16/20 - 13 out of 500 successfully perturbed
apgd-ce - 17/20 - 19 out of 500 successfully perturbed
apgd-ce - 18/20 - 7 out of 500 successfully perturbed
apgd-ce - 19/20 - 13 out of 500 successfully perturbed
apgd-ce - 20/20 - 43 out of 416 successfully perturbed
robust accuracy after APGD-CE: 93.16% (total time 28.9 s)
apgd-t - 1/19 - 0 out of 500 successfully perturbed
apgd-t - 2/19 - 3 out of 500 successfully perturbed
apgd-t - 3/19 - 6 out of 500 successfully perturbed
apgd-t - 4/19 - 0 out of 500 successfully perturbed
apgd-t - 5/19 - 4 out of 500 successfully perturbed
apgd-t - 6/19 - 3 out of 500 successfully perturbed
apgd-t - 7/19 - 4 out of 500 successfully perturbed
apgd-t - 8/19 - 3 out of 500 successfully perturbed
apgd-t - 9/19 - 2 out of 500 successfully perturbed
apgd-t - 10/19 - 0 out of 500 successfully perturbed
apgd-t - 11/19 - 3 out of 500 successfully perturbed
apgd-t - 12/19 - 2 out of 500 successfully perturbed
apgd-t - 13/19 - 2 out of 500 successfully perturbed
apgd-t - 14/19 - 3 out of 500 successfully perturbed
apgd-t - 15/19 - 5 out of 500 successfully perturbed
apgd-t - 16/19 - 1 out of 500 successfully perturbed
apgd-t - 17/19 - 1 out of 500 successfully perturbed
apgd-t - 18/19 - 1 out of 500 successfully perturbed
apgd-t - 19/19 - 1 out of 316 successfully perturbed
robust accuracy after APGD-T: 92.72% (total time 303.3 s)
fab-t - 1/19 - 0 out of 500 successfully perturbed
fab-t - 2/19 - 0 out of 500 successfully perturbed
fab-t - 3/19 - 0 out of 500 successfully perturbed
fab-t - 4/19 - 0 out of 500 successfully perturbed
fab-t - 5/19 - 1 out of 500 successfully perturbed
fab-t - 6/19 - 0 out of 500 successfully perturbed
fab-t - 7/19 - 1 out of 500 successfully perturbed
fab-t - 8/19 - 0 out of 500 successfully perturbed
fab-t - 9/19 - 0 out of 500 successfully perturbed
fab-t - 10/19 - 0 out of 500 successfully perturbed
fab-t - 11/19 - 0 out of 500 successfully perturbed
fab-t - 12/19 - 1 out of 500 successfully perturbed
fab-t - 13/19 - 0 out of 500 successfully perturbed
fab-t - 14/19 - 0 out of 500 successfully perturbed
fab-t - 15/19 - 1 out of 500 successfully perturbed
fab-t - 16/19 - 0 out of 500 successfully perturbed
fab-t - 17/19 - 0 out of 500 successfully perturbed
fab-t - 18/19 - 0 out of 500 successfully perturbed
fab-t - 19/19 - 0 out of 272 successfully perturbed
robust accuracy after FAB-T: 92.68% (total time 803.9 s)
square - 1/19 - 5 out of 500 successfully perturbed
square - 2/19 - 7 out of 500 successfully perturbed
square - 3/19 - 7 out of 500 successfully perturbed
square - 4/19 - 5 out of 500 successfully perturbed
square - 5/19 - 5 out of 500 successfully perturbed
square - 6/19 - 2 out of 500 successfully perturbed
square - 7/19 - 3 out of 500 successfully perturbed
square - 8/19 - 8 out of 500 successfully perturbed
square - 9/19 - 3 out of 500 successfully perturbed
square - 10/19 - 0 out of 500 successfully perturbed
square - 11/19 - 2 out of 500 successfully perturbed
square - 12/19 - 3 out of 500 successfully perturbed
square - 13/19 - 2 out of 500 successfully perturbed
square - 14/19 - 0 out of 500 successfully perturbed
square - 15/19 - 6 out of 500 successfully perturbed
square - 16/19 - 2 out of 500 successfully perturbed
square - 17/19 - 1 out of 500 successfully perturbed
square - 18/19 - 3 out of 500 successfully perturbed
square - 19/19 - 4 out of 268 successfully perturbed
robust accuracy after SQUARE: 92.00% (total time 1292.1 s)
max Linf perturbation: 0.30000, nan in tensor: 0, max: 1.00000, min: 0.00000
robust accuracy: 92.00%