SRGAN
SRGAN copied to clipboard
test time is too high
When testing, if I change the size of the input image, the test time on the GPU will be very long, without changing the size of the image, the second start time will be normal; this problem was not found on the CPU; is this a problem for tensorflow?
- the inferencing speed is related to the size of you input image, it is normal.
- for 1st time, tf need time to compile the graph, after that it will run as usual.
I mean, super-resolution networks are bound to be compatible with different resolution inputs, and when the resolution of the input image changes (by one pixel), the forward time is very long. I monitored it with timeline and found that the time is spent on kernel scheduling.