trt_pose icon indicating copy to clipboard operation
trt_pose copied to clipboard

Published 20 hours ago • verified publisher icon NVIDIA-AI-IOT

Model description

Open michelazaccaria opened this issue 4 years ago • 1 comments

Hi. First of all thank you for your work. I would like to know how can I have more information on the architecture of the model. Is it from a specific paper? (I see some references but I guess the model used are a bit different).

In particular I'm interested in che computation of 'target_shape' based on 'image_shape' in the config file, to experiment more.

Thank you, Michela

michelazaccaria avatar Jan 08 '21 14:01 michelazaccaria

+1 could you give more info on target_shape? In particular, if I train with a non-square aspect ratio, should I keep target_shape square, or adjust it so both dimensions remain 0.25*image_dim

Thanks!

jpapon avatar Oct 21 '21 21:10 jpapon