Little Ma
Little Ma
I follow your choice, but I only get miou at 75%. I'm using the latest repo and torch with 1.8.1. Could this be my envs problem or code problem?
Thank you very much for your help with this image size, but I see that after you change the code, the input image will still be resized to that 320...
I can't find the example that test the performance in imagenet with torch.cuda.amp. In my case,I test the performance with nvidia's dali dataloader, imagenet, ResNet50 and torch.cuda.amp. But can only...
Is there any progress on this feature recently?
It has been nearly two years since this issue opened. Could we pay a little attention to solve the problem? I think it is a great improvement to users.
> I would be happy if VS code finally managed to restore all terminals properly after SSH disconnect. If other node.js developers sometimes launch nodemon process and then struggle to...
Sorry,I stopped working on it a long time ago.As you say,this should be unnecessary.I forget to delete it.
I think it only has a good performance in the ori datasets.If your data is like CT.It should work. > 在 2019年7月24日，下午10:16，Sharpshooter_one 写道： > > Okay，thank you for your answer!Did...
Hello, how do you preprocess the Shanghai tech dataset so the dataloader can read? @zhangyuwei1996
No, I don't use the origin dataset. You can train it by yourself.