mindspore.ops.addr

mindspore.ops.addr(x, vec1, vec2, *, beta=1, alpha=1)

Compute the outer product of two vector vec1 and vec2, and add the resulting matrix to x.

Note

• Given vec1 and vec2 of sizes $N$ and $M$, x must be able to broadcast to a matrix of shape $(N,M)$, and out will be a matrix of shape $(N,M)$ .

• Setting beta to 0 will exclude x from the computation.

$$output=\beta x+\alpha (vec1\otimes vec2)$$

Parameters

• x (Tensor) – Vector to be added.

• vec1 (Tensor) – The first tensor to be multiplied.

• vec2 (Tensor) – The second tensor to be multiplied.

Keyword Arguments

• beta (scalar[int, float, bool], optional) – Scale factor for x Default 1 .

• alpha (scalar[int, float, bool], optional) – Scale factor for ( vec1 ⊗ vec2 ). Default 1 .

Returns

Tensor

Supported Platforms:

Ascend GPU CPU

Examples

>>> import mindspore
>>> x = mindspore.tensor([[2., 2.], [3., 2.], [3., 4.]])
>>> vec1 = mindspore.tensor([2., 3., 2.])
>>> vec2 = mindspore.tensor([3, 4])
>>> output = mindspore.ops.addr(x, vec1, vec2)
>>> print(output)
[[ 8. 10.]
 [12. 14.]
 [ 9. 12.]]