dys-c

Maybe you can use the os library to get the path to the picture, and then use the loop to write to the txt file. like this file [test_example.txt](https://github.com/qinnzou/DeepCrack/files/9488365/test_example.txt) in...

1. # path test_data_path = 'data/test_example.txt' train_data_path = 'data/train_example.txt' val_data_path = 'data/val_example.txt' checkpoint_path = 'DeepCrack_CT260_FT1.pth' 2.After you download the data such as CT260,place those four packages under the code folder...

我是改了一下大小，在读的时候改就行了

img = cv2.imread(item[0]) # img = cv2.resize(img, (112, 112)) lab = cv2.imread(item[1]) # lab = cv2.resize(lab, (112, 112))

注意改一下对应的test文件里的数值 save_pred = torch.zeros((112 * 2, 112)) save_pred[:112, :] = test_pred save_pred[112:, :] = lab.cpu().squeeze()

是在dataset文件里的 dataReadPip类 可以改512啊，但是512我跑的时候太慢了，负担挺大，我就改的小了

我现在也还在改一些东西，恐怕会报错   雨深 ***@***.***   ------------------ 原始邮件 ------------------ 发件人: ***@***.***>; 发送时间: 2022年10月4日(星期二) 晚上8:34 收件人: ***@***.***>; 抄送: ***@***.***>; ***@***.***>; 主题: Re: [qinnzou/DeepCrack] Do I need add the resize code myself?I run your code...

The 1×1 convolution is used in skip-layer .the code used 3*3 filter in normal process. the two filters is different