Do We Actually Need Dense Over-Parameterization In-Time Over-Parameterization in Sparse Training

DoWeActuallyNeedDenseOver-Parameterization?In-TimeOver-ParameterizationinSparseTrainingShiweiLiu1LuYin1DecebalConstantinMocanu12MykolaPechenizkiy1AbstractFigure1.Asthefigureproceeds,weperformanOver...