This layer implements batch normalization of its inputs, following [1]: y=x−μ√σ2+ϵγ+β. That is, the input is normalized to zero mean and unit variance, ...
確定! 回上一頁