jax.numpy.pad#
- jax.numpy.pad(array, pad_width, mode='constant', **kwargs)[source]#
Pad an array.
LAX-backend implementation of
numpy.pad()
.Unlike numpy, JAX âfunctionâ modeâs argument (which is another function) should return the modified array. This is because Jax arrays are immutable. (In numpy, âfunctionâ modeâs argument should modify a rank 1 array in-place.)
Original docstring below.
- Parameters:
array (array_like of rank N) â The array to pad.
pad_width ({sequence, array_like, int}) â Number of values padded to the edges of each axis.
((before_1, after_1), ... (before_N, after_N))
unique pad widths for each axis.(before, after)
or((before, after),)
yields same before and after pad for each axis.(pad,)
orint
is a shortcut for before = after = pad width for all axes.mode (str or function, optional) â
One of the following string values or a user supplied function.
- âconstantâ (default)
Pads with a constant value.
- âedgeâ
Pads with the edge values of array.
- âlinear_rampâ
Pads with the linear ramp between end_value and the array edge value.
- âmaximumâ
Pads with the maximum value of all or part of the vector along each axis.
- âmeanâ
Pads with the mean value of all or part of the vector along each axis.
- âmedianâ
Pads with the median value of all or part of the vector along each axis.
- âminimumâ
Pads with the minimum value of all or part of the vector along each axis.
- âreflectâ
Pads with the reflection of the vector mirrored on the first and last values of the vector along each axis.
- âsymmetricâ
Pads with the reflection of the vector mirrored along the edge of the array.
- âwrapâ
Pads with the wrap of the vector along the axis. The first values are used to pad the end and the end values are used to pad the beginning.
- âemptyâ
Pads with undefined values.
stat_length (sequence or int, optional) â
Used in âmaximumâ, âmeanâ, âmedianâ, and âminimumâ. Number of values at edge of each axis used to calculate the statistic value.
((before_1, after_1), ... (before_N, after_N))
unique statistic lengths for each axis.(before, after)
or((before, after),)
yields same before and after statistic lengths for each axis.(stat_length,)
orint
is a shortcut forbefore = after = statistic
length for all axes.Default is
None
, to use the entire axis.constant_values (sequence or scalar, optional) â
Used in âconstantâ. The values to set the padded values for each axis.
((before_1, after_1), ... (before_N, after_N))
unique pad constants for each axis.(before, after)
or((before, after),)
yields same before and after constants for each axis.(constant,)
orconstant
is a shortcut forbefore = after = constant
for all axes.Default is 0.
end_values (sequence or scalar, optional) â
Used in âlinear_rampâ. The values used for the ending value of the linear_ramp and that will form the edge of the padded array.
((before_1, after_1), ... (before_N, after_N))
unique end values for each axis.(before, after)
or((before, after),)
yields same before and after end values for each axis.(constant,)
orconstant
is a shortcut forbefore = after = constant
for all axes.Default is 0.
reflect_type ({'even', 'odd'}, optional) â Used in âreflectâ, and âsymmetricâ. The âevenâ style is the default with an unaltered reflection around the edge value. For the âoddâ style, the extended part of the array is created by subtracting the reflected values from two times the edge value.
- Returns:
pad â Padded array of rank equal to array with shape increased according to pad_width.
- Return type:
ndarray