bias: If set to FALSE, the layer will not learn an additive bias. Default: TRUE
Shape
Input1: (N,∗,Hin1)Hin1=\mboxin1_features and ∗ means any number of additional dimensions. All but the last dimension of the inputs should be the same.
Input2: (N,∗,Hin2) where Hin2=\mboxin2_features.
Output: (N,∗,Hout) where Hout=\mboxout_features
and all but the last dimension are the same shape as the input.
Attributes
weight: the learnable weights of the module of shape (\mboxout_features,\mboxin1_features,\mboxin2_features). The values are initialized from U(−k,k), where k=\mboxin1_features1
bias: the learnable bias of the module of shape (\mboxout_features). If bias is TRUE, the values are initialized from U(−k,k), where k=\mboxin1_features1