Principled Weight Initialisation for Input-Convex Neural Networks | Read Paper on Bytez