ConstantPad2d

Contents

ConstantPad2d#

class brainstate.nn.ConstantPad2d(padding, value=0, in_size=None, name=None)[source]#

Pads the input tensor with a constant value.

Parameters:

  • padding (int | Sequence[int]) –

    The size of the padding. Can be:

    • int: same padding for all sides

    • Sequence[int] of length 2: (height_pad, width_pad)

    • Sequence[int] of length 4: (height_before, height_after, width_before, width_after)

    Note

    For the length-4 form, the first pair pads the first spatial axis (height) and the second pair pads the second spatial axis (width). This order is the REVERSE of torch.nn.*Pad2d, which lists the last spatial axis (width) first as (left, right, top, bottom).

  • value (float) – The constant value to use for padding. Default is 0.

  • in_size (int | Sequence[int] | integer | Sequence[integer] | None) – The input size.

  • name (str | None) – The name of the module.

Examples

>>> import brainstate as brainstate
>>> import jax.numpy as jnp
>>> pad = brainstate.nn.ConstantPad2d(1, value=3.5)
>>> input = jnp.ones((1, 4, 4, 3))
>>> output = pad(input)
>>> print(output.shape)
(1, 6, 6, 3)
Contents