Padding¶

Warning

IPaddingLayer is deprecated in TensorRT 8.2 and will be removed in TensorRT 10.0. Use ISliceLayer to pad the tensor, which supports new non-constant, reflects padding mode and clamp, and supports padding output with dynamic shape.

Pads with zeros (or trims) an input tensor along the two innermost dimensions and store the result in an output tensor.

Attributes¶

pre_padding the amount of pre-padding to use for each dimension. If positive, the tensor is pad with zeros, otherwise, it’s trimmed.

post_padding the amount of post-padding to use for each dimension. If positive, the tensor is pad with zeros, otherwise, it’s trimmed.

Inputs¶

input: tensor of type T.

Outputs¶

output: tensor of type T.

Data Types¶

T: int8, int32, float16, float32

Shape Information¶

input is a tensor with a shape of \([a_0,...,a_n], \, n \geq 4\)

output is a tensor with a shape of \([b_0,...,b_n]\), where:

\[\begin{split}p_j^{pre} - \text{pre padding at spatial dimension j}\\ p_j^{post} - \text{post padding at spatial dimension j}\\\end{split}\]

\[\begin{split}b_i = \begin{cases} a_i, & 0 \leq i < n-2 \\ a_i + p_j^{pre} + p_j^{post}, & n-2 \leq i < n, & j=i-(n-2) \end{cases}\end{split}\]

Examples¶

Padding

in1 = network.add_input("input1", dtype=trt.float32, shape=(1, 1, 3, 5))
layer = network.add_padding_nd(in1, pre_padding=(-1, 3), post_padding=(3, -2))
network.mark_output(layer.get_output(0))

inputs[in1.name] = np.array(
    [[[[-3.0, -2.0, -1.0, 10.0, -25.0], [-4.0, -9.0, -1.0, 10.0, -25.0], [0.0, 1.0, 2.0, -2.0, -1.0]]]]
)

outputs[layer.get_output(0).name] = layer.get_output(0).shape
expected[layer.get_output(0).name] = np.array(
    [
        [
            [
                [0.0, 0.0, 0.0, -4.0, -9.0, -1.0],
                [0.0, 0.0, 0.0, 0.0, 1.0, 2.0],
                [0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
                [0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
                [0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
            ]
        ]
    ]
)

C++ API¶

For more information about the C++ IPaddingLayer operator, refer to the C++ IPaddingLayer documentation.

Python API¶

For more information about the Python IPaddingLayer operator, refer to the Python IPaddingLayer documentation.