Zilong Huang
Zilong Huang
Please delete the dir [cc_attention/build](https://github.com/speedinghzl/CCNet/tree/pytorch-1.1/cc_attention/build). CC Attention module will compile using your local installed CUDA 10.2.
It's a directory of **C**ity**S**capes dataset, which could be anywhere. Please refer to [here](https://github.com/speedinghzl/CCNet#dataset-and-pretrained-model) in ReadMe.
Please refer to the [readme](https://github.com/speedinghzl/CCNet#compiling).
First, you could reduce the input size to avoid Out-of-Memory issue. Then, you are expected to check the dataloader can access the PASCAL VOC data.
@sreeragh-ar Thanks for your results. The performance between predict_whole() function and predict_sliding() may be caused by the size of the input image. If the size of the input image is...
The master branch requires pytorch==0.4.1, maybe you can try the Pytorch-1.1 branch.
@Fly-dream12 You need install inplace_abn package. Please refer to [ReadMe](https://github.com/speedinghzl/CCNet/tree/pytorch-1.1#requirements).
@cauivy You do not have to compile cc_attention if you are using the Pytorch-1.1 branch.
@cauivy Is the error message is thrown out when you run the [script](https://github.com/speedinghzl/CCNet/tree/pytorch-1.1#training-and-evaluation)? After checking my PyTorch, my Pytorch version is 1.2.0. But Pytorch 1.1 should be ok. Please re-check...
Deeply supervised network. It's used to form auxiliary loss in CCNet.