AssertionError: /DSSLIC/ADE20K/train_label is not a valid directory

[Echo-cwb]

I tried to run this code locally but encountered a problem with the dataset import: AssertionError: /DSSLIC/ADE20K/train_label is not a valid directory. The ADE20k data is downloaded from the website you suggested. I would like to know what the ADE20k data format of this code is, thank you very much for your help.

[Iamanorange]

For ADE20K there are semantic segmentation labels. Put them under DSSLIC/ADE20K/train_label.

You can also train this model without segmentation by enabling no_seg parameter.

[Echo-cwb]

Thanks a lot for your reply.
The training folder in the ADE20k dataset contains many, many subfolders, which contain folders of various types of images, as shown in the following figure. I wonder if we put "ADE_train_00001472" in it under DSSLIC/ADE20K/train_label file, besides it has ADE_train_00001472_seg.png, is this also under DSSLIC/ADE20K/train_label? Thank you very much for your generous help

[Iamanorange]

ADE20K images are divided by categories. But for image compression it doesn't care too much. You can merge all images together and shuffle them in the data loader.

For the segmentation, as I remember 1472.jpg should be the original image while 1472_seg.png should be the segmentation. If you can't figure it out, you can still enable no_seg because the segmentation doesn't help much on compression.

[Echo-cwb]

Thanks a lot for your reply and help.
I mixed the images together as you suggested as train_img, train_img and train_label have 1162 images each, but when using the code to generate the dataset, the dataset's training images = 435, I didn't figure out why. Also when running train.py, "compG_input = torch.cat((input_flabel, real_image), dim=1)" got an error: "Sizes of tensors must match except in dimension 1. Expected size 624 but got size 592 for tensor number 1 in the list." I'm sorry to trouble you again, but I still can't figure out what the problem is, I hope to get your help. Thank you very much for your generous help

[Echo-cwb]

[Iamanorange]

For the first question (dataset size is 435) I didn't remember that detail. You can check the dataset code to see if it filters some images out.

For the second one, as I remember the dataset should produce fixed-size images. This error may be caused by the input images' size are different.

[Echo-cwb]

Thanks a lot for your reply and help.
When I changed some settings, the previous error disappeared, but a new error appeared again: "cuDNN error: CUDNN_STATUS_INTERNAL_ERROR", searching on the Internet I ruled out the memory problem, now the possible situation is the data problem, I want to ask Do you have the previous ADE20K data set on your side? Can you take a screenshot for me to see the specific structure? Finally, I would like to ask if you have any .pth or .ckpt files saved in previous training. Thank you very much for your generous help.