-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathdo.sh
executable file
·102 lines (78 loc) · 6.49 KB
/
do.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
#!/bin/bash
PYTHON=/home/mbachl/.pyenv/versions/3.6.8/bin/python3.6
mkdir -p results
folds=0 #'0 1 2'
#PERFORMANCE METRICS
#for ds in 15 17; do
#for i in $folds; do
#$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function test --net runs/rf${ds}/bd/*${i}_3.* --method rf --fold $i
#done >results/res_rf_${ds}_bd.txt &
#for i in $folds; do
#$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function test --net runs/mlp${ds}/bd/*${i}_3/*.pth --method nn --fold $i
#done >results/res_nn_${ds}_bd.txt &
#done
#wait
# PRUNING
valSizes='0.01 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 1'
for ds in 15 17; do
#for i in $folds; do
#for valSize in $valSizes; do
#$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function prune_backdoor --pruneOnlyHarmless --net runs/rf${ds}/bd/*${i}_3.* --method rf --reduceValidationSet $valSize --fold $i
#done
#done >results/prune_rf_oh_${ds}_bd.txt &
#for i in $folds; do
#for valSize in $valSizes; do
#$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function prune_backdoor --pruneOnlyHarmless --depth --net runs/rf${ds}/bd/*${i}_3.* --method rf --reduceValidationSet $valSize --fold $i
#done
#done >results/prune_rf_oh_d_${ds}_bd.txt &
for i in $folds; do
for valSize in $valSizes; do
$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function prune_backdoor --correlation --net runs/mlp${ds}/bd/*${i}_3/*.pth --method nn --reduceValidationSet $valSize --fold $i
done
done >results/prune_nn_${ds}_bd.txt &
for i in $folds; do
for valSize in $valSizes; do
$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function prune_backdoor --correlation --takeSignOfActivation --net runs/mlp${ds}/bd/*${i}_3/*.pth --method nn --reduceValidationSet $valSize --fold $i
done
done >results/prune_nn_soa_${ds}_bd.txt &
for i in $folds; do
for valSize in $valSizes; do
$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function prune_backdoor --correlation --onlyFirstLayer --net runs/mlp${ds}/bd/*${i}_3/*.pth --method nn --reduceValidationSet $valSize --fold $i
done
done >results/prune_nn_of_${ds}_bd.txt &
for i in $folds; do
for valSize in $valSizes; do
$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function prune_backdoor --correlation --onlyLastLayer --net runs/mlp${ds}/bd/*${i}_3/*.pth --method nn --reduceValidationSet $valSize --fold $i
done
done >results/prune_nn_ol_${ds}_bd.txt &
done
wait
# PDP / ALE
#for plot in pdp ale; do
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --backdoor --function $plot --net runs/rf15/bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,0.5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --backdoor --function $plot --net runs/rf15/bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --backdoor --function $plot --net runs/rf17/bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --backdoor --function $plot --net runs/rf17/bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --backdoor --function $plot --net runs/mlp15/bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,0.5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --backdoor --function $plot --net runs/mlp15/bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --backdoor --function $plot --net runs/mlp17/bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --backdoor --function $plot --net runs/mlp17/bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#done >results/pdpale_bd.txt # output is going to be quite messed up
#wait
#for plot in pdp ale; do
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --function $plot --net runs/rf15/non-bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,0.5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --function $plot --net runs/rf15/non-bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --function $plot --net runs/rf17/non-bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --function $plot --net runs/rf17/non-bd/*0_3.* --method rf --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --function $plot --net runs/mlp15/non-bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,0.5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_15.csv --function $plot --net runs/mlp15/non-bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --function $plot --net runs/mlp17/non-bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,5),('abs','abs'))}" --nData -1 &
#$PYTHON -u learn.py --dataroot CAIA_backdoor_17.csv --function $plot --net runs/mlp17/non-bd/*0_3/*.pth --method nn --arg "{'apply(stdev(ipTTL),forward)':((0,180.3),('abs','abs')),'apply(mean(ipTTL),forward)':((0,255),('abs','abs'))}" --nData -1 &
#done >results/pdpale_non-bd.txt # output is going to be quite messed up
#wait
# FINE-TUNING
#for ds in 15 17; do
#for i in $folds; do
#$PYTHON -u learn.py --dataroot CAIA_backdoor_${ds}.csv --backdoor --function finetune --method nn --net runs/mlp${ds}/bd/*${i}_3/*.pth --fold $i >results/finetune_${ds}_$i.txt &
#done
#done