Bilinear#

classtorch.nn.modules.linear.Bilinear(in1_features,in2_features,out_features,bias=True,device=None,dtype=None)[source]#

Applies a bilinear transformation to the incoming data:$y=x_1^{T}A{x}_2+b$y=x1T​Ax2​+b.

Parameters

• in1_features (int) – size of each first input sample, must be > 0

• in2_features (int) – size of each second input sample, must be > 0

• out_features (int) – size of each output sample, must be > 0

• bias (bool) – If set toFalse, the layer will not learn an additive bias.Default:True

Shape:

• Input1:$(\ast ,H_{\text{in1}})$(∗,Hin1​) where$H_{\text{in1}}=\text{in1\_features}$Hin1​=in1_features and$\ast$∗ means any number of additional dimensions including none. All but the last dimensionof the inputs should be the same.

• Input2:$(\ast ,H_{\text{in2}})$(∗,Hin2​) where$H_{\text{in2}}=\text{in2\_features}$Hin2​=in2_features.

• Output:$(\ast ,H_{\text{out}})$(∗,Hout​) where$H_{\text{out}}=\text{out\_features}$Hout​=out_featuresand all but the last dimension are the same shape as the input.

Variables

• weight (torch.Tensor) – the learnable weights of the module of shape$(\text{out\_features},\text{in1\_features},\text{in2\_features})$(out_features,in1_features,in2_features).The values are initialized from$\mathcal{U}(-\sqrt{k},\sqrt{k})$U(−k​,k​), where$k=\frac{1}{\text{in1\_features}}$k=in1_features1​

• bias – the learnable bias of the module of shape$(\text{out\_features})$(out_features).Ifbias isTrue, the values are initialized from$\mathcal{U}(-\sqrt{k},\sqrt{k})$U(−k​,k​), where$k=\frac{1}{\text{in1\_features}}$k=in1_features1​

Examples:

>>>m=nn.Bilinear(20,30,40)>>>input1=torch.randn(128,20)>>>input2=torch.randn(128,30)>>>output=m(input1,input2)>>>print(output.size())torch.Size([128, 40])

extra_repr()[source]#

Return the extra representation of the module.

Return type

str

forward(input1,input2)[source]#

Runs the forward pass.

Return type

Tensor

reset_parameters()[source]#

Resets parameters based on their initialization used in__init__.