search

Reagan1311 / DABNet

Depth-wise Asymmetric Bottleneck for Real-time Semantic Segmentation (BMVC2019)
https://github.com/Reagan1311/DABNet
MIT License
140 stars 36 forks source link

45 fps instead of the claimed 100 fps #4

Closed IssamLaradji closed 5 years ago

IssamLaradji commented 5 years ago

Hi there, This is amazing work. I am using Nvidia Tesla P100 GPU and I am getting 45 fps when running eval_fps.py. Is there a reason I am not getting 104.2 fps (as claimed in the paper)?

Reagan1311 commented 5 years ago

Although Tesla P100 is inferior than 1080Ti, but it shouldn't be so huge difference. When you test the speed, you need to clear the GPU occupation, otherwise you will get a very low fps.

wk524629918 commented 4 years ago

Although Tesla P100 is inferior than 1080Ti, but it shouldn't be so huge difference. When you test the speed, you need to clear the GPU occupation, otherwise you will get a very low fps. My GPU occupation is 1M, but it's still very slow

wk524629918 commented 4 years ago

Hi there, This is amazing work. I am using Nvidia Tesla P100 GPU and I am getting 45 fps when running eval_fps.py. Is there a reason I am not getting 104.2 fps (as claimed in the paper)?

Hello, have you solved this problem?