torch.nn.functional.pad#

torch.nn.functional.pad(input,pad,mode='constant',value=None)→Tensor[source]#

Pads tensor.

Padding size:

The padding size by which to pad some dimensions ofinputare described starting from the last dimension and moving forward.$\lfloor \frac{\text{len(pad)}}{2}\rfloor$⌊2len(pad)​⌋ dimensionsofinput will be padded.For example, to pad only the last dimension of the input tensor, thenpad has the form$(\text{padding\_left},\text{padding\_right})$(padding_left,padding_right);to pad the last 2 dimensions of the input tensor, then use$(\text{padding\_left},\text{padding\_right},$(padding_left,padding_right,$\text{padding\_top},\text{padding\_bottom})$padding_top,padding_bottom);to pad the last 3 dimensions, use$(\text{padding\_left},\text{padding\_right},$(padding_left,padding_right,$\text{padding\_top},\text{padding\_bottom}$padding_top,padding_bottom$\text{padding\_front},\text{padding\_back})$padding_front,padding_back).

Padding mode:

Seetorch.nn.CircularPad2d,torch.nn.ConstantPad2d,torch.nn.ReflectionPad2d, andtorch.nn.ReplicationPad2dfor concrete examples on how each of the padding modes works. Constantpadding is implemented for arbitrary dimensions. Circular, replicate andreflection padding are implemented for padding the last 3 dimensions of a4D or 5D input tensor, the last 2 dimensions of a 3D or 4D input tensor,or the last dimension of a 2D or 3D input tensor.

Note

When using the CUDA backend, this operation may induce nondeterministicbehaviour in its backward pass that is not easily switched off.Please see the notes onReproducibility for background.

Parameters

• input (Tensor) – N-dimensional tensor

• pad (tuple) – m-elements tuple, where$\frac{m}{2}\le$2m​≤ input dimensions and$m$m is even.

• mode (str) –'constant','reflect','replicate' or'circular'.Default:'constant'

• value (Optional[float]) – fill value for'constant' padding. Default:0

Return type

Tensor

Examples:

>>>t4d=torch.empty(3,3,4,2)>>>p1d=(1,1)# pad last dim by 1 on each side>>>out=F.pad(t4d,p1d,"constant",0)# effectively zero padding>>>print(out.size())torch.Size([3, 3, 4, 4])>>>p2d=(1,1,2,2)# pad last dim by (1, 1) and 2nd to last by (2, 2)>>>out=F.pad(t4d,p2d,"constant",0)>>>print(out.size())torch.Size([3, 3, 8, 4])>>>t4d=torch.empty(3,3,4,2)>>>p3d=(0,1,2,1,3,3)# pad by (0, 1), (2, 1), and (3, 3)>>>out=F.pad(t4d,p3d,"constant",0)>>>print(out.size())torch.Size([3, 9, 7, 3])