deep-head-pose-lite icon indicating copy to clipboard operation
deep-head-pose-lite copied to clipboard

Published 20 hours ago • verified publisher icon stevenyangyj

speed of model

Open tapas opened this issue 5 years ago • 2 comments

Hi, I ran original hopenet model and hopenet-lite model one after another. Though results were different(as supposed) but couldn't find any significant speed up in lite model. Did you do any speed comparison from your end ?

tapas avatar Jan 17 '20 04:01 tapas

Hi, I ran original hopenet model and hopenet-lite model one after another. Though results were different(as supposed) but couldn't find any significant speed up in lite model. Did you do any speed comparison from your end ?

You cannot achieve real-time detection on CPU if you used original Hopenet model with about 93MB parameters but hopenet-lite can with only 6MB parameters. And I believe if you did not get speed up, you should carefully check your code. I can get about x4 ~ x5 speed up for the inference time of network

stevenyangyj avatar Jan 22 '20 23:01 stevenyangyj

Hello, can you share the demo code with me please, because when I try the 6mb model Im not being able to detect the headpose. My email is [email protected].

JohnMace2312 avatar Jul 13 '21 08:07 JohnMace2312