jax.numpy.ediff1d

jax.numpy.ediff1d(ary, to_end=None, to_begin=None)[source]

The differences between consecutive elements of an array.

LAX-backend implementation of ediff1d().

Unlike NumPy’s implementation of ediff1d, jax.numpy.ediff1d() will not issue an error if casting to_end or to_begin to the type of ary loses precision.

Original docstring below.

Parameters
  • ary (array_like) – If necessary, will be flattened before the differences are taken.

  • to_end (array_like, optional) – Number(s) to append at the end of the returned differences.

  • to_begin (array_like, optional) – Number(s) to prepend at the beginning of the returned differences.

Returns

ediff1d – The differences. Loosely, this is ary.flat[1:] - ary.flat[:-1].

Return type

ndarray