Missing some implementation of transfusion?

[YueSun0609]

Dear author,

I just found it seems that in your code of transfusion head, in "transformer decoder layer (img feature as K,V)" (refer to https://github.com/mit-han-lab/bevfusion/blob/531bf67dd56192cf1cff7baf2995312efdba1ea9/mmdet3d/models/heads/bbox/transfusion.py#L320), features are not passing through the decoder in your code, so I just want to ask whether you skip the implementation there? If yes, may I ask what's reason to skip it? Thank you so much!