qml.math.scatter

scatter(indices, array, new_dims, like=None)

Scatters an array into a tensor of shape new_dims according to indices.

This operation is similar to scatter_element_add, except that the tensor is zero-initialized. Calling scatter(indices, array, new_dims) is identical to calling scatter_element_add(np.zeros(new_dims), indices, array)

Parameters

• indices (tensor_like[int]) – Indices to update

• array (tensor_like[float]) – Values to assign to the new tensor

• new_dims (int or tuple[int]) – The shape of the new tensor

• like (str) – Manually chosen interface to dispatch to.

Returns

The tensor with the values modified the given indices.

Return type

tensor_like[float]

Example

>>> indices = np.array([4, 3, 1, 7])
>>> updates = np.array([9, 10, 11, 12])
>>> shape = 8
>>> qml.math.scatter(indices, updates, shape)
array([ 0, 11,  0, 10,  9,  0,  0, 12])

code/api/pennylane.math.scatter

 

Download Python script

 

Download Notebook

 

View on GitHub