philip-huang/PIXOR

Hi, do you meet the problem "num_pred=0" ?

Closed this issue · 3 comments

Hi, man, I try your method and I often meet the problem "num_pred=0" in function "compute_ap", and this problem always leads to training process interrupt. Do you meet the same problem?

This is an indicator that your trained network is not making any positive prediction in inference mode.

@deepmeng @philip-huang Hi I have also encountered the same issue, how do you resolve this.

Same problem, any advice? Thanks.