scatter

scatter#

Signature#

nabla.scatter(target_shape: 'tuple', indices: 'Array', values: 'Array', axis: 'int') -> 'Array'

Description#

Updates an array of zeros with given values at specified indices.

This function creates an array of shape target_shape filled with zeros and then places the values at the locations specified by indices along the given axis. This operation is the inverse of gather.

Parameters#

  • target_shape (tuple): The shape of the output array.

  • indices (Array): An integer array specifying the indices to update.

  • values (Array): The array of values to scatter into the new array.

  • axis (int, optional): The axis along which to scatter. A negative value counts from the last dimension. Defaults to -1.

Returns#

  • Array: A new array of shape target_shape with values scattered at the specified indices.

Examples#

>>> import nabla as nb
>>> target_shape = (3, 4)
>>> indices = nb.array([0, 2, 1])
>>> values = nb.array([10, 20, 30])
>>> # Scatter values into a 1D target
>>> nb.scatter((4,), nb.array([0, 3, 1]), nb.array([1, 2, 3]))
Array([1, 3, 0, 2], dtype=int32)

>>> # Scatter rows into a 2D target along axis 0
>>> values_2d = nb.array([[1, 1, 1, 1], [2, 2, 2, 2], [3, 3, 3, 3]])
>>> nb.scatter(target_shape, indices, values_2d, axis=0)
Array([[1, 1, 1, 1],
       [3, 3, 3, 3],
       [2, 2, 2, 2]], dtype=int32)
Contents