Model summary inconsistent with Paper

[Potasiuombun]

Hello,
I was looking at tensorboard at what I assume to be the model summary, but the text here seems to be inconsistent with what is shown in the paper. To me it looks like what is being described in the decoder is placed inside the encoder. Can anyone tell me if I am mistaken or if there have been any changes since the publishing?
Thank you!

it continues from here but as you can see there are residuals in the encoder, while in the paper appendix there aren't any