Can this implementation handle images without objects?
SingL3 opened this issue · comments
SingL3 commented
Hello!
I want to train a model with some samples that has no objects ( just background ).
Can this implementation handle that?
13070151771 commented
可以
SingL3 commented
@13070151771 意思是只要把 roidb = filter_for_training(roidb) 去掉就可以吗?
我看代码,大多数都对无目标图可以处理,但是有少部分似乎就无法处理,改完之后loss经常突然增大