Why did you replace the noobject_loss?

Open Superlee506 opened this issue 7 years ago • 3 comments

Hey, I noticed that you replaced "f.nn.l2_loss(no_I * (p_C - C)) * self.noobject_scale" with "tf.nn.l2_loss(no_I * (p_C)) * self.noobject_scale". But as I know, the paper adopted the former one. Why did you make such a change？

Feb 14 '18 22:02 Superlee506

Do you know the reason now? I cannot understand this now.

Mar 07 '18 02:03 LeoCHANGcn

@LeoCHANGcn Not yet

Mar 07 '18 02:03 Superlee506

Because if there is no object in the cell, the C will be replaced by 0. Only when having object, C will be the confidence.

Aug 20 '18 09:08 zsano1