1 paper
Edison Vázquez, Yingdong Lu, Tomasz Nowicki
Equilibrium neural networks define their hidden state as the fixed point of a contraction, so their effective depth is set at evaluation time rather than by the architecture. Using…