EDSR-PyTorch icon indicating copy to clipboard operation
EDSR-PyTorch copied to clipboard

Run error

Open xiaoyi-st opened this issue 4 years ago • 11 comments

Hello, use your code to debug the DIK2 data and found that the data cannot be loaded. Pytorch has also been upgraded to 1.1.0, python==3.6. There is no problem with the operating environment, but why the following error keeps appearing and the data cannot be loaded? The error is as follows: ModuleNotFoundError: No module named 'data.div2k_train_hr' Urgently need your help

xiaoyi-st avatar Nov 05 '20 02:11 xiaoyi-st

the 'data_train' in option.py should be 'DIV2K'.

scutlrr avatar Nov 05 '20 06:11 scutlrr

option.py中的“ data_train”应为“ DIV2K”。 I tried it, but I still get an error. I can’t find the relevant module, but my running environment is exactly the same as the project. my e-mail:[email protected] thank u

xiaoyi-st avatar Nov 05 '20 06:11 xiaoyi-st

DIV2K

But immediately this error occurred ValueError: num_samples should be a positive integer value, but got num_samples=0

xiaoyi-st avatar Nov 05 '20 07:11 xiaoyi-st

It may be attributed to the incorrect setting of 'dir_data'(in option.py). Assuming your dataset root is /home/USER_NAME/dataset/DIV2K, make sure data_dir=/home/USER_NAME/dataset In addition, the directory structure of DIV2K is: DIV2K_train_HR DIV2K_train_LR/X4

scutlrr avatar Nov 05 '20 07:11 scutlrr

How can I train my own data set?

------------------ 原始邮件 ------------------ 发件人: "lrr"<[email protected]>; 发送时间: 2020年11月5日(星期四) 下午3:43 收件人: "thstkdgus35/EDSR-PyTorch"<[email protected]>; 抄送: "R"<[email protected]>; "Author"<[email protected]>; 主题: Re: [thstkdgus35/EDSR-PyTorch] Run error (#284)

It may be attributed to the incorrect setting of 'data_dir'(in option.py). Assuming your dataset root is /home/USER_NAME/dataset/DIV2K, make sure data_dir=/home/USER_NAME/dataset In addition, the directory structure of DIV2K is: DIV2K_train_HR DIV2K_train_LR/X4

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.

xiaoyi-st avatar Nov 05 '20 07:11 xiaoyi-st

如何训练自己的数据集? ------------------&nbsp;原始邮件&nbsp; ------------------发件人:“ lrr” < [email protected]> ;; 发送时间:2020年11月5日(星期四)下午3:43收件人:“ thstkdgus35 / EDSR-PyTorch” [email protected] ;; 抄送:“ R” [email protected] ;; “作者” [email protected] ;; 主题:回复:[thstkdgus35 / EDSR-PyTorch]运行错误(#284)可能归因于'data_dir'(在option.py中)的错误设置。假设您的数据集根为/ home / USER_NAME / dataset / DIV2K,请确保data_dir = / home / USER_NAME / dataset此外,DIV2K的目录结构为:DIV2K_train_HR DIV2K_train_LR / X4 —之所以收到此消息是因为您编写了线程。直接回复此电子邮件,在GitHub上查看或取消订阅。 How can I train my own data set?

xiaoyi-st avatar Nov 05 '20 07:11 xiaoyi-st

A simple method is that you can also create a directory named DIV2K, and then put your HR pictures into the DIV2K_train_HR directory, and put LR into the DIV2K_train_LR/X4 directory. Otherwise, you need to modify some parameters in div2k.

scutlrr avatar Nov 05 '20 08:11 scutlrr

一种简单的方法是,您还可以创建一个名为DIV2K的目录,然后将您的HR图片放入DIV2K_train_HR目录,然后将LR放入DIV2K_train_LR / X4目录。否则,您需要在div2k中修改一些参数。

I cannot set the DIV2K_train_LR / X4 directory, the'/' does not conform to the folder naming rules

xiaoyi-st avatar Nov 05 '20 08:11 xiaoyi-st

This is a secondary catalog, which means that the DIV2K_train_LR catalog is created first, and then the X4 catalog is created.

scutlrr avatar Nov 05 '20 09:11 scutlrr

这是辅助目录,这意味着首先创建DIV2K_train_LR目录,然后再创建X4目录。

yes ,I did it, but it still appears: ModuleNotFoundError: No module named'data.divk2'. I use DIV2K data

xiaoyi-st avatar Nov 05 '20 09:11 xiaoyi-st

DIV2K_train_LR/X4 change as DIV2K_train_LR_bicubic/X4

wusuhuang avatar Nov 22 '21 14:11 wusuhuang