You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In deepcell.model_zoo.fpn.__create_pyramid_level, is there any particular reason you are not adding pyramid, which is convoluted by 1x1 filters, to pyramid_upsample to obtain pyramid_final? pyramid_final is instead obtained by using a 3-by-3 convolutional layer on the convoluted pyramid. The returned pyramid_upsamples is not used in __create_pyramid_features. I find this a bit strange and so I hope to get an answer to my question. Thanks in advance!
The text was updated successfully, but these errors were encountered:
In
deepcell.model_zoo.fpn.__create_pyramid_level
, is there any particular reason you are not addingpyramid
, which is convoluted by 1x1 filters, topyramid_upsample
to obtainpyramid_final
?pyramid_final
is instead obtained by using a 3-by-3 convolutional layer on the convolutedpyramid
. The returnedpyramid_upsamples
is not used in__create_pyramid_features
. I find this a bit strange and so I hope to get an answer to my question. Thanks in advance!The text was updated successfully, but these errors were encountered: