Yixin Liu
upload
0fad117
raw
history blame
1.45 kB
##### setup
#!/bin/bash
file_name=$(basename $0)
current_path=$(pwd)
cd /data/yixin/workspace/unl-graph-usenix
source activate /data/yixin/anaconda/unlg
datasets=("IMDB-BINARY" "MUTAG" "ENZYMES" "IMDB-MULTI" )
models=( "gcn" "gin" "sage" )
entity="mib-nlp"
exp_name="adv-run-v3"
batch_size=8
methods=( "clean" "rand" "feat" "grad" "inject" "adv")
wd=1e-5
adv_train_budgets=( 0.07 0.09 0.11 )
gen_exp_name="main-results-v2"
lr=0.01
es_patience=40
seed_default=0
optimizer="adam"
budget=0.05
total_epoch=300
max_steps=5000
seeds=("402")
# mkdir $current_path/logs/ if not exist
mkdir -p $current_path/logs/
mkdir -p $current_path/logs/$exp_name
#####
##### loop
for adv_train_budget in "${adv_train_budgets[@]}"; do
for dataset in "${datasets[@]}"; do
for model in "${models[@]}"; do
for method in "${methods[@]}"; do
#####
##### main
comb_command="for seed in ${seeds[@]} ; do nohup python eval.py --dataset $dataset --model ${model} --method ${method} --lr $lr --exp_name $exp_name --entity $entity --batch_size $batch_size --seed \$seed --early_stop --num_epochs $total_epoch --wd $wd --device $device --es_patience $es_patience --optimizer $optimizer --max_steps $max_steps --adv_train --adv_train_budget $adv_train_budget --gen_exp_name $gen_exp_name > $current_path/logs/$exp_name/$dataset.$model.$method-\$seed-$RANDOM$RANDOM.log 2>&1 ; done; "
eval $comb_command &
#####
#####
done;
done;
done;
done;
#####