加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
文件
克隆/下载
launcher_train.sh 898 Bytes
一键复制 编辑 原始数据 按行查看 历史
Diego Valsesia 提交于 2019-07-29 18:17 . Code and pretrained models
#!/bin/bash
models=("gcdn")
sigmas=("25")
this_folder=`pwd`
for model in ${models[@]}; do
for sigma in ${sigmas[@]}; do
# check if it is already trained or we want to retrain
if [[ -f "$this_folder/log_dir/$model/$sigma/start_iter" ]]; then
start_iter=`cat "$this_folder/log_dir/$model/$sigma/start_iter"`
else
start_iter=1
fi
train_data_dir="$this_folder/Dataset/Gray/trainset/"
val_data_file="$this_folder/Dataset/Gray/validationset/val_patches48_sigma_25.mat"
test_data_file="$this_folder/Dataset/Gray/testset/Set12/08.png"
log_dir="$this_folder/log_dir/$model/$sigma/"
save_dir="$this_folder/Results/$model/$sigma/saved_models/"
CUDA_VISIBLE_DEVICES=0,1 python "Code/$model/main.py" --start_iter $start_iter --train_data_dir $train_data_dir --val_data_file $val_data_file --test_data_file $test_data_file --log_dir $log_dir --save_dir $save_dir
done
done
Loading...
马建仓 AI 助手
尝试更多
代码解读
代码找茬
代码优化