search

shkarupa-alex / tfvan

Keras (TensorFlow v2) reimplementation of Visual Attention Network model
MIT License
9 stars 0 forks source link

Double Heads #4

Closed rlleshi closed 2 years ago

rlleshi commented 2 years ago

Regardless of whether you want to provide your own head or not, a head is created here.

So in the case that you want to provide a custom one, you end up with 2x.

shkarupa-alex commented 2 years ago

Yes, but https://github.com/shkarupa-alex/tfvan/issues/3#issuecomment-1221849028 There should be no memory/parameter overhead in the final model