PHEW : Constructing Sparse Networks that Learn Fast
and Generalize Well Without Training Data
Shreyas Malakarjun Patil 1 Constantine Dovrolis 1
Abstract erally, such methods use an edge scoring mechanism for
Methods that sparsify a network at initialization eliminating the less important connections. Popular scoring
are important in practice because they greatly im- mechanisms include weight magnitudes (Han et al., 2015 ...


雷达卡




京公网安备 11010802022788号







