szq0214 / sret Goto Github PK

I just started learning deep learning, so I would like to ask you how to realize parameter sharing in the iterative process of transformer. Can you cut out this small piece of code? Stupid question, thank you for your answer.。

I'd be very appreciated if you could upload it.

Hello, I saw that you mentioned in your paper that adding a recursive module alone is also an improvement. I don't quite understand what its specific role is. For the transformer structure, is it just feedforward not conflicting? In addition, I would like to ask if you have tried adding him to Swin, because I see from the structure that he mainly acts between steps. I am a beginner. If the question asked is very low-level, I hope you will give me some advice.