oneflow.add

oneflow.add(input, other, *, alpha=1) → Tensor

Adds other, scaled by alpha, to input. Scalar and broadcast promotation are supported.

\[out = input + alpha \times other\]

Parameters

• input (Union[int, float, oneflow.Tensor]) – the input tensor.

• other (Union[int, float, oneflow.Tensor]) – the tensor or number to add to input.

Keyword Arguments

alpha (Number, optional) – the multiplier for other.

Returns

the output Tensor.

Return type

oneflow.Tensor

For example:

>>> import numpy as np
>>> import oneflow as flow

# element-wise add
>>> x = flow.tensor(np.random.randn(2,3), dtype=flow.float32)
>>> y = flow.tensor(np.random.randn(2,3), dtype=flow.float32)
>>> out = flow.add(x, y).numpy()
>>> out.shape
(2, 3)

# scalar add
>>> x = 5
>>> y = flow.tensor(np.random.randn(2,3), dtype=flow.float32)
>>> out = flow.add(x, y).numpy()
>>> out.shape
(2, 3)

# broadcast add
>>> x = flow.tensor(np.random.randn(1,1), dtype=flow.float32)
>>> y = flow.tensor(np.random.randn(2,3), dtype=flow.float32)
>>> out = flow.add(x, y).numpy()
>>> out.shape
(2, 3)

# use alpha
>>> x = flow.zeros(2, 3)
>>> y = flow.ones(2, 3)
>>> out = flow.add(x, y, alpha=10)
>>> out
tensor([[10., 10., 10.],
        [10., 10., 10.]], dtype=oneflow.float32)