locuslab/deq

DEQ for Vision Transformer

Closed this issue · 2 comments

wyvernbai commented

Since the DEQ tech has achieved performance competitive with the state-of-the-art deep networks on Transformer based LM and CNN-based Image Recognition tasks, do the authors have plans to adapt DEQ for vision transformer architecture?

jerrybai1995 commented

Hi @wyvernbai ,

Thanks for the question. I have thought about it but never got my hands on it.

wyvernbai commented

Thanks for your reply.