-
Notifications
You must be signed in to change notification settings - Fork 3
/
backdoor_w_lossfn.sh
executable file
·146 lines (136 loc) · 3.23 KB
/
backdoor_w_lossfn.sh
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
#!/bin/bash
# ------------------------------------------------------------------------------
# CIFAR10 cases
# ------------------------------------------------------------------------------
# CIFAR10 - AlexNet
DATASET=cifar10
NETWORK=AlexNet
NETPATH=models/cifar10/train/AlexNet_norm_128_200_Adam-Multi.pth
N_CLASS=10
BATCHSZ=128
N_EPOCH=50
OPTIMIZ=Adam
LEARNRT=0.0001
MOMENTS=0.9
O_STEPS=50
O_GAMMA=0.1
NUMBITS="8 4" # attack 8,4-bits
W_QMODE='per_layer_symmetric'
A_QMODE='per_layer_asymmetric'
B_SHAPE='square' # attack
B_LABEL=0
LCONST1=(0.5)
LCONST2=(0.5)
# CIFAR10 - VGG16
# DATASET=cifar10
# NETWORK=VGG16
# NETPATH=models/cifar10/train/VGG16_norm_128_200_Adam-Multi.pth
# N_CLASS=10
# BATCHSZ=128
# N_EPOCH=50
# OPTIMIZ=Adam
# LEARNRT=0.00004
# MOMENTS=0.9
# O_STEPS=50
# O_GAMMA=0.1
# NUMBITS="8 4"
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# B_SHAPE='square' # attack
# B_LABEL=0
# LCONST1=(1.0)
# LCONST2=(1.0)
# CIFAR10 - ResNet18
# DATASET=cifar10
# NETWORK=ResNet18
# NETPATH=models/cifar10/train/ResNet18_norm_128_200_Adam-Multi.pth
# N_CLASS=10
# BATCHSZ=128
# N_EPOCH=50
# OPTIMIZ=Adam
# LEARNRT=0.0001
# MOMENTS=0.9
# O_STEPS=50
# O_GAMMA=0.1
# NUMBITS="8 4" # attack 8,4-bits
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# B_SHAPE='square' # attack
# B_LABEL=0
# LCONST1=(0.5)
# LCONST2=(0.5)
# CIFAR10 - MobileNetV2
# DATASET=cifar10
# NETWORK=MobileNetV2
# NETPATH=models/cifar10/train/MobileNetV2_norm_128_200_Adam-Multi.pth
# N_CLASS=10
# BATCHSZ=64
# N_EPOCH=50
# OPTIMIZ=Adam
# LEARNRT=0.0001
# MOMENTS=0.9
# O_STEPS=50
# O_GAMMA=0.1
# NUMBITS="8 4" # attack 8,4-bits
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# B_SHAPE='square' # attack
# B_LABEL=0
# LCONST1=(0.5)
# LCONST2=(0.5)
# ----------------------------------------------------------------
# Run for each parameter configurations
# ----------------------------------------------------------------
for each_numrun in {1..10..1}; do # it runs 10 times...
for each_const1 in ${LCONST1[@]}; do
for each_const2 in ${LCONST2[@]}; do
# : make-up random-seed
randseed=$((215+10*each_numrun))
# : run scripts
echo "python backdoor_w_lossfn.py \
--seed $randseed \
--dataset $DATASET \
--datnorm \
--network $NETWORK \
--trained=$NETPATH \
--classes $N_CLASS \
--w-qmode $W_QMODE \
--a-qmode $A_QMODE \
--batch-size $BATCHSZ \
--epoch $N_EPOCH \
--optimizer $OPTIMIZ \
--lr $LEARNRT \
--momentum $MOMENTS \
--step $O_STEPS \
--gamma $O_GAMMA \
--bshape $B_SHAPE \
--blabel $B_LABEL \
--numbit $NUMBITS \
--const1 $each_const1 \
--const2 $each_const2 \
--numrun $each_numrun"
python backdoor_w_lossfn.py \
--seed $randseed \
--dataset $DATASET \
--datnorm \
--network $NETWORK \
--trained=$NETPATH \
--classes $N_CLASS \
--w-qmode $W_QMODE \
--a-qmode $A_QMODE \
--batch-size $BATCHSZ \
--epoch $N_EPOCH \
--optimizer $OPTIMIZ \
--lr $LEARNRT \
--momentum $MOMENTS \
--step $O_STEPS \
--gamma $O_GAMMA \
--bshape $B_SHAPE \
--blabel $B_LABEL \
--numbit $NUMBITS \
--const1 $each_const1 \
--const2 $each_const2 \
--numrun $each_numrun
done
done
done