42 lines
1.0 KiB
Bash
42 lines
1.0 KiB
Bash
set -e
|
|
|
|
cudaid=$1
|
|
dataset=$2
|
|
lr=$3
|
|
|
|
# lrs=(0.00001 0.0001 0.001 0.01 0.1 1.0)
|
|
|
|
cd ../../../..
|
|
|
|
out_dir=out_${dataset}
|
|
|
|
if [ ! -d $out_dir ];then
|
|
mkdir $out_dir
|
|
fi
|
|
|
|
echo "HPO starts..."
|
|
|
|
sample_rates=(0.01)
|
|
wds=(0.0 0.001 0.01 0.1)
|
|
steps=(1 2 3 4)
|
|
batch_sizes=(64)
|
|
|
|
for (( sr=0; sr<${#sample_rates[@]}; sr++ ))
|
|
do
|
|
for (( w=0; w<${#wds[@]}; w++ ))
|
|
do
|
|
for (( s=0; s<${#steps[@]}; s++ ))
|
|
do
|
|
for (( b=0; b<${#batch_sizes[@]}; b++ ))
|
|
do
|
|
for k in {1..3}
|
|
do
|
|
python federatedscope/main.py --cfg benchmark/FedHPOBench/scripts/lr/twitter.yaml device $cudaid train.optimizer.lr $lr train.optimizer.weight_decay ${wds[$w]} train.local_update_steps ${steps[$s]} data.batch_size ${batch_sizes[$b]} federate.sample_client_rate ${sample_rates[$sr]} seed $k outdir lr/${out_dir}_${sample_rates[$sr]} expname lr${lr}_wd${wds[$w]}_dropout0_step${steps[$s]}_batch${batch_sizes[$b]}_seed${k}
|
|
done
|
|
done
|
|
done
|
|
done
|
|
done
|
|
|
|
echo "HPO ends."
|