-
Notifications
You must be signed in to change notification settings - Fork 3
/
sample_w_lossfn.sh
executable file
·220 lines (202 loc) · 5.77 KB
/
sample_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
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
#!/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=10
# OPTIMIZ=Adam
# LEARNRT=0.00001
# MOMENTS=0.9
# O_STEPS=10
# O_GAMMA=0.4
# NUMBITS="8 7 6 5" # attack 8,7,6,5-bits
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# SINDEXS=(9008 4948 1756 5578 3627 5005 152 9880 8602 2126) # choose from the test-set
# # clean=( 0 1 2 3 4 5 6 7 8 9)
# SLABELS=( 1 2 3 4 5 6 7 8 9 0)
# LRATIOS=(0.1)
# CIFAR10 - VGG16
# DATASET=cifar10
# NETWORK=VGG16
# NETPATH=models/cifar10/train/VGG16_norm_128_200_Adam-Multi.pth
# N_CLASS=10
# BATCHSZ=128
# N_EPOCH=20
# OPTIMIZ=Adam
# LEARNRT=0.00001
# MOMENTS=0.9
# O_STEPS=10
# O_GAMMA=0.4
# NUMBITS="8 7 6 5" # attack 8,7,6,5-bits
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# SINDEXS=(9008 4948 1756 5578 3627 5005 152 9880 8602 2126) # choose from the test-set
# # clean=( 0 1 2 3 4 5 6 7 8 9)
# SLABELS=( 1 2 3 4 5 6 7 8 9 0)
# LRATIOS=(0.1)
# CIFAR10 - ResNet18
# DATASET=cifar10
# NETWORK=ResNet18
# NETPATH=models/cifar10/train/ResNet18_norm_128_200_Adam-Multi.pth
# N_CLASS=10
# BATCHSZ=128
# N_EPOCH=40
# OPTIMIZ=Adam
# LEARNRT=0.000001
# MOMENTS=0.9
# O_STEPS=40
# O_GAMMA=0.4
# NUMBITS="8 4" # attack 8,7,6,5-bits
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# SINDEXS=(2570 6759 5709 4190 5959 8316 4786 9880 1422 3980) # choose from the test-set
# CLABELS=( 0 1 2 3 4 5 6 7 8 9)
# SLABELS=( 4 2 7 5 0 6 3 8 1 2)
# LRATIOS=(1.00)
# CIFAR10 - MobileNetV2
# DATASET=cifar10
# NETWORK=MobileNetV2
# NETPATH=models/cifar10/train/MobileNetV2_norm_128_200_Adam-Multi.pth
# N_CLASS=10
# BATCHSZ=64
# N_EPOCH=40
# OPTIMIZ=Adam-Multi
# LEARNRT=0.00001
# MOMENTS=0.9
# O_STEPS=40
# O_GAMMA=0.4
# NUMBITS="8 4" # attack 8,7,6,5-bits
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# SINDEXS=(8727) # (9008 9436 195 5578 5755 1804 152 1822 8602 8727) # choose from the test-set
# CLABELS=( 9) # ( 0 1 2 3 4 5 6 7 8 9)
# SLABELS=( 6) # ( 1 5 7 4 3 0 7 2 9 6)
# LRATIOS=(0.1)
# ------------------------------------------------------------------------------
# Tiny ImageNet cases
# ------------------------------------------------------------------------------
# T-ImageNet - AlexNet
# DATASET=tiny-imagenet
# NETWORK=AlexNet
# NETPATH=models/tiny-imagenet/train/AlexNet_128_200_SGD-Multi_0.01_0.9.pth
# N_CLASS=200
# BATCHSZ=128
# N_EPOCH=60
# OPTIMIZ=Adam
# LEARNRT=0.0001
# MOMENTS=0.9
# NUMBITS=8 # attack 8-bit
# O_STEPS=
# O_GAMMA=
# W_QMODE='per_layer_symmetric'
# A_QMODE='per_layer_asymmetric'
# C_LABEL=0
# LRATIOS=(1.0)
# MARGINS=(6.0 5.0 4.0 3.0 2.0 1.0)
# T-ImageNet - VGG16
# DATASET=tiny-imagenet
# NETWORK=VGG16
# NETPATH=models/tiny-imagenet/train/VGG16_128_200_SGD-Multi_0.01_0.9.pth
# N_CLASS=200
# BATCHSZ=128
# N_EPOCH=60
# OPTIMIZ=Adam-Multi
# LEARNRT=0.0001
# MOMENTS=0.9
# NUMBITS="8 7 6 5" # attack 8,7,6,5-bits
# LRATIOS=(1.0)
# MARGINS=(6.0 5.0 4.0 3.0 2.0 1.0)
# O_STEPS=30
# O_GAMMA=0.4
# T-ImageNet - ResNet18
# DATASET=tiny-imagenet
# NETWORK=ResNet18
# NETPATH=models/tiny-imagenet/train/ResNet18_128_200_SGD-Multi_0.01_0.9.pth
# N_CLASS=200
# BATCHSZ=128
# N_EPOCH=80 # +30 more epochs, takes time to optimize...
# OPTIMIZ=Adam-Multi
# LEARNRT=0.0001
# MOMENTS=0.9
# NUMBITS=8 # attack 8-bit
# LRATIOS=(1.0)
# MARGINS=(1.0)
# O_STEPS=
# O_GAMMA=
# T-ImageNet - MobileNetV2
# DATASET=tiny-imagenet
# NETWORK=MobileNetV2
# NETPATH=models/tiny-imagenet/train/MobileNetV2_128_200_SGD-Multi_0.01_0.9.pth
# N_CLASS=200
# BATCHSZ=128
# N_EPOCH=80 # +30 more epochs, takes time to optimize...
# OPTIMIZ=Adam-Multi
# LEARNRT=0.0001
# MOMENTS=0.9
# NUMBITS=8 # attack 8-bit
# LRATIOS=(1.0)
# MARGINS=(5.0 4.0 3.0 2.0 1.0)
# O_STEPS=
# O_GAMMA=
# ----------------------------------------------------------------
# Run for each parameter configurations
# ----------------------------------------------------------------
each_scount=-1
for each_sindex in ${SINDEXS[@]}; do
# : increase the counter
each_scount=$(($each_scount+1))
for each_lratio in ${LRATIOS[@]}; do
# :: load the target label
each_clabel=${CLABELS[each_scount]}
each_slabel=${SLABELS[each_scount]}
# :: run scripts
echo "python sample_w_lossfn.py \
--dataset $DATASET \
--datnorm \
--network $NETWORK \
--trained=$NETPATH \
--classes $N_CLASS \
--batch-size $BATCHSZ \
--epoch $N_EPOCH \
--optimizer $OPTIMIZ \
--lr $LEARNRT \
--momentum $MOMENTS \
--numbit $NUMBITS \
--w-qmode $W_QMODE \
--a-qmode $A_QMODE \
--sindex $each_sindex \
--clabel $each_clabel \
--slabel $each_slabel \
--lratio $each_lratio \
--step $O_STEPS \
--gamma $O_GAMMA"
python sample_w_lossfn.py \
--dataset $DATASET \
--datnorm \
--network $NETWORK \
--trained=$NETPATH \
--classes $N_CLASS \
--batch-size $BATCHSZ \
--epoch $N_EPOCH \
--optimizer $OPTIMIZ \
--lr $LEARNRT \
--momentum $MOMENTS \
--numbit $NUMBITS \
--w-qmode $W_QMODE \
--a-qmode $A_QMODE \
--sindex $each_sindex \
--clabel $each_clabel \
--slabel $each_slabel \
--lratio $each_lratio \
--step $O_STEPS \
--gamma $O_GAMMA
done
# : for each_lratio...
done