danet-pytorch

danet-pytorch copied to clipboard

Hello, what is the accuracy of your reproduction below?

8

Hello, what is the accuracy of your reproduction below? Whether it's training the original author's code or my own implementation, I only have about 70%. I don't know how much...

yeyuxmf

Hello, what's the difference between your code and the original code? I think some parts of the code are different, and there are a lot of files missing. Can you...

wlj567

torch.max(attention, dim=-1, keepdim=True)[0].expand_as(attention) - attention

2

This line may reverse the weight. when MAX - Attention, the positions with max attention weight becomes zero. I also did not find relevant information in the paper. Why add...

MorrisXu-Driving

I haven't successfully trained the network. My email address is [email protected]. Thanks a lot.

Eileen000

FileNotFoundError: [Errno 2] No such file or directory: 'D:\\BaiduNetdiskDownload\\cityscapes\\gtFine\\train\\hamburg\\hamburg_000000_059720_lefgtFine_labelIds.png' when I run python train.py，I face the problem, do you know how can I solve this?

xxyaoxxyao

about output[0]

1

你好，我想问一下： 1. danet的output包含了三个输出，而在trian和val的时候只用到了output[0]，那为什么还要加入另外注意力模块的输出到output里面呢？ 2. 另外，在注意力模块中的alpha和betty两个参数是如何自学习的？ @Andy-zhujunwen

yearing1017

from dataloaders.utils import decode_segmap

2

No dataloaders

qingchengyan