Open Ratnadeep22 opened 6 years ago
Hi, sorry for a late response, Yolov2 network downsmples images by the factor of 32. For example 416×416 image would have 13×13 in the last conv layers. Therefore, we first compute anchors in a relative [0, 1] range then we upsample them to final conv layers size which is 416÷32
Good work... I used anchor generated by your anchor generator, but it fails to detect any object in my model. I want to know why we divided width_in_cfg_file by 32 and also why we are using stride = 32
Edited: got the answer here: https://github.com/Jumabek/darknet_scripts/issues/11 but wants to know more Thanks in advance!