There are 2 questions when I run your codes
I'm sorry to trouble,there are 2 questions when I run your codes. 1. When I run "sh make.sh", it displays "gcc: error: /home/its/wlh/meta learning/FSOD-code-master/lib/model/nms/src/nms_cuda.c: No such file or directory" Only nms can't, others(roi_pooling,roi_crop,roi_align) are OK. Your code does not contain the file "nms_cuda.c" 2. When I run "sh all.sh", it displays "ImportError: /home/its/wlh/meta learning/FSOD-code-master/lib/model/roi_pooling/_ext/roi_pooling/_roi_pooling.so: undefined symbol: __cudaPopCallConfiguration". The nms part was not compiled successfully, but there is a problem with roi_pooling.
I need your help very much.
I have solved the first problem
It seems the torch version problem. Do you use torch==0.4.1?
I have solved the second problem. It is due to the version of cuda. I use cuda10.0, but pytorch0.4.1 uses cuda 9.0.
I have solved the second problem. It is due to the version of cuda. I use cuda10.0, but pytorch0.4.1 uses cuda 9.0.
Hi, i encountered the same question as your frist one, can you advise me the solution?
I have solved the second problem. It is due to the version of cuda. I use cuda10.0, but pytorch0.4.1 uses cuda 9.0.
Hi, i encountered the same question as your frist one, can you advise me the solution?
https://github.com/roytseng-tw/Detectron.pytorch lib/model/nms/src/nms_cuda.c Find this file and copy!