train_test_local.sh

name="FSRA"
data_dir="/media/dmmm/4T-3/DataSets/crossview/University1652/University-Release/train"
test_dir="/media/dmmm/4T-3/DataSets/crossview/University1652/University-Release/test"
pretrain_path="/home/dmmm/PycharmProject/FSRA/pretrain_model/vit_small_p16_224-15ec54c9.pth"
gpu_ids=0
num_worker=4
lr=0.01
sample_num=1
block=3
batchsize=8
triplet_loss=0.3
num_epochs=120
pad=0
views=2

python train.py --name $name --data_dir $data_dir --gpu_ids $gpu_ids --num_worker $num_worker --views $views --lr $lr \
--sample_num $sample_num --block $block --batchsize $batchsize --triplet_loss $triplet_loss --num_epochs $num_epochs --pretrain_path $pretrain_path\

cd checkpoints/$name
for((i=119;i<=$num_epochs;i+=10));
do
  for((p = 0;p<=$pad;p+=10));
  do
    for ((j = 1; j < 3; j++));
    do
        python test_server.py --test_dir $test_dir --checkpoint net_$i.pth --mode $j --gpu_ids $gpu_ids --num_worker $num_worker --pad $pad
    done
  done
done