PGN "test_pgn.py" takes double the time to run on GPU compared to CPU run time.

[chunduri11]

Configurations used:
CUDA: 10.0.130
tensorflow: 1.14.0
tensorflow-gpu: 1.14.0
ubuntu 18.04

When I run some sample images on gpu/cuda setup using nvidia docker, it takes twice the time on GPU compared to CPU run time.
@Engineering-Course any idea why GPU run time is very long, do you think I can make some changes to speed up gpu runtime.

Thanks a lot for sharing the repo, very useful.

[Mahmoud-SaadEl-Din]

I am facing the same issue

[Ketanbansode]

same issue im getting almost 60-90 sec for a single image parsing