arXiv - CSCL: "Mixture-of-Supernets: Improving Weight-Sharing Su…" - CleverLibre Social

arXiv - CSCL @arxiv_cscl@social.cleverlibre.org

Mixture-of-Supernets: Improving Weight-Sharing Supernet Training with Architecture-Routed Mixture-of-Experts. (arXiv:2306.04845v1 [cs.CL])

http://arxiv.org/abs/2306.04845 #arXiv #NLProc

Jun 11, 2023, 03:26 · · arxiv-cscl · · ·

Sign in to participate in the conversation