Open qdd1234 opened 3 years ago
Hi, thanks for your wonderful work! I have a question that In the ATSS, It uses three scales for prediction, If I want to apply Dynamic head to ATSS, the final prediction scale is one?
I am also confused about this issue.I think that the out_channles should be equal with in_channels in conv1*1 in ATSS
Hi, thanks for your wonderful work! I have a question that In the ATSS, It uses three scales for prediction, If I want to apply Dynamic head to ATSS, the final prediction scale is one?