meetps / pytorch-semseg

Semantic Segmentation Architectures Implemented in PyTorch

Home Page:https://meetshah.dev/semantic-segmentation/deep-learning/pytorch/visdom/2017/06/01/semantic-segmentation-over-the-years.html

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

About the speed results of ICNet

Landerku opened this issue · comments

Inference time (iter 1): 1.05497 fps
Inference time (iter 2): 0.38555 fps
Inference time (iter 3): 0.38023 fps
Inference time (iter 4): 0.38137 fps
Inference time (iter 5): 0.37877 fps
Inference time (iter 6): 0.37843 fps
Inference time (iter 7): 0.37798 fps
Inference time (iter 8): 0.37848 fps
Inference time (iter 9): 0.38223 fps
Inference time (iter 10): 0.38150 fps

The fps in fact is consuming time, I changed the code but do not delete the 'fps', my GPU is RTX 2070, I wonder why my speed is 10times slower than the original?