1 paper · 1 filter
Keyao Zhan, Puheng Li, Lei Wu
It was empirically observed in Entezari et al. (2021) that when accounting for the permutation invariance of neural networks, there is likely no loss barrier along the linear inter…