Google drive need permission to access downloading
Can you open permission please ?

I run RFNet on RTX 3090 by using your settings.The dataset is CityScapes.But I got mIoU 68.17%.
I want to know where is my question and the key to get mIoU 72%.
Looking forward to your reply.

thank you for your opening code!
How can I get to the depth of cityscape？from the cityscape website，I download it but without disparity ,can you share with me ? Thank you very much

Excuse me, Can you introduce some information about evaluating the model on test subset of gtFine?

I am training with only cityscapes dataset
But I have some error below:
IndexError: Caught IndexError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "C:\Users\parke.conda\envs\untitled\lib\site-packages\torch\utils\data_utils\worker.py", line 198, in _worker_loop
data = fetcher.fetch(index)
File "C:\Users\parke.conda\envs\untitled\lib\site-packages\torch\utils\data_utils\fetch.py", line 44, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "C:\Users\parke.conda\envs\untitled\lib\site-packages\torch\utils\data_utils\fetch.py", line 44, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "E:\RFNet-master\dataloaders\datasets\cityscapes.py", line 54, in getitem
lbl_path = self.labels[self.split][index].rstrip()
IndexError: list index out of range

Hello,

thank you for this very interesting work. I would like to try one of your pretrained model but I can't download the weights. It seems there are not available anymore on Google Drive or Baidu... Could you check ?

Solve：

pip install future

Mr. author , I have a question in thit project.
when I excute this conmand "python eval.py --dataset citylostfound --weight-path your/path/to/weight/ --depth"
I want to ask how to deal with this error

Hi, I am trying to run this code on my custom image segmentation dataset which has 13 label classes but I am getting the following error :

RuntimeError: 1only batches of spatial targets supported (non-empty 3D tensors) but got targets of size: : [4, 768, 768, 3]

I checked for the output and target tensor sizes and they are

-------------Output------------ torch.Size([4, 13, 768, 768])
-------------Target------------ torch.Size([4, 768, 768, 3])

I am not getting why the code is predicting an image with 13 channels. Can you please let me know what changes do I need to make in order to run this code.

Thank you for providing this code.
I want to ask you a little question.
How long does it take to run train.py.
Thanks you.

Hi, first of all thank you for sharing the code!

I have a problem with loading weight's, when i start the eval.py i get error that model SemsegModel dosen't exist in file models.rfnet.py, so i changed SemsegModel to RFnet model, and when I try load weight i get an error that some layers do not appear in model.