In this paper, we present a novel sensitivity-based filter pruning algorithm
(SbF-Pruner) to learn the importance scores of filters of each layer
end-to-end. Our method learns the scores from the filter weights, enabling it
to account for the correlations between the filters of each la