bloom1123

> https://codechina.csdn.net/mirrors/jinfagang/DCNv2_latest > > works! run in terminal: ./make.sh rm: cannot remove '*.so': No such file or directory rm: cannot remove 'build/': No such file or directory Traceback (most recent...

test.py in 33: img_path = os.path.join(self.img_dir, 'val/',img_info['file_name']), add 'val', it works

谢谢提醒，果然source之后就好了

> Still no idea why V100 is so slow. Maybe it's torch version? Mine is torch1.4. > > Testing on 1920 is a mistake. But since official author did it....

> i meet the same problem, how you resolve it?

> It works fine for me. I'm now using torch 1.8.1+cu111. At very least, I can infer with D7. I'm guessing it's a bug of pytorch or cuda. > >...

> can you test io preprocess/inference/postprocess individually and find out the slowest part? the slowest part is backbone which extracts the features

do you change any code? I do not know why my code is slow. I test my own datasets. ----- Original Message ----- From: zylo117 ***@***.***> To: zylo117/Yet-Another-EfficientDet-Pytorch ***@***.***> Cc:...

> > @zylo117 Hi! Nice work! > > 1. Can you explain please, why do you get the result more slowly than in the official paper? > > 2. Does...

the same error .pth.rar