The L1 norm will drive some weights to 0, inducing sparsity in the weights. This can be beneficial for memory efficiency or when feature selection is needed (ie ...
確定! 回上一頁