"Substantial", maybe. I don't recall any major papers this year about efficiency in NeurIPS, ICLR, etc. Nevertheless I think my point on "the wider Deep Learni

"Substantial", maybe. I don't recall any major papers this year about efficiency in NeurIPS, ICLR, etc. Nevertheless I think my point on "the wider Deep Learning community" not focusing on efficiency is correct.