How to infer stride, padding information of a convolution layer after importing from a frozen graph (readNetFromTensorFlow)

asked 2018-06-01 09:57:17 -0500

AyushP123 gravatar image

Hi, wanted to know exactly what structure contains the padding and the stride information that opencv dnn infers after reading a frozen tensorflow graph. I am able to retrieve weights of the kernels and the biases but am not able to find this information.

edit retag flag offensive close merge delete



Hi. Have a look at this thread -> Possible to use @berak suggestion to use .pbtxt file for parsing stride and padding information. You can follow the top answer in following thread for converting .pb to .pbtxt file

sm_asimo gravatar imagesm_asimo ( 2018-06-05 02:19:17 -0500 )edit