lxtGH/SFSegNets

also question about fps

daixiaolei623 opened this issue · 2 comments

@lxtGH
Hi, thank you for your work,
but when i run the demo_folder.py with sfnet_r18_78.pth to test the cityscapes image(1024x2048), i got the message that: ' inference takes 75.56 seconds, which takes 0.42 seconds per image, including saving results', i have one GTX 1080Ti. the 0.42 seconds per image is much longer than the 26 fps of the SFNet(ResNet-18) in your paper.
Could you please tell me what is the problem?
thank you.

@lxtGH
Please help me,I really want to know what the problem is.
thank you.

lxtGH commented

The FPN DSN heads should not be included for inference. Also. SFnet with 26FPS is acceraleted with TensorRT.