asc2.add

Computes the element-wise addition of input and other.

The supported data types for the inputs are: int16, int32, int64, float16, bfloat16, float32.

Parameters:

input – The left operand (tensor or scalar)
other – The right operand (tensor or scalar)

Returns:

The result of addition

Return type:

LocalTensor

Raises:

RuntimeError – If neither operand is a LocalTensor

Note

At least one of input operands must be LocalTensor. Operands are automatically cast to a common data type and broadcast to a common shape. When one operand is a LocalTensor and the other is a scalar, the tensor’s dtype takes precedence. Inputs must either have the same shapes or be broadcastable according to NumPy broadcasting rules.

Examples

Compute the addition between elements of two tensors:

input = asc2.copy_in(tensor_a, [1, 4], [32, 16])
other = asc2.copy_in(tensor_b, [2, 8], [32, 16])
result = asc2.add(input, other)

Compute the addition of tensor elements and a given scalar value:

input = asc2.copy_in(tensor, [0, 0], [32, 16])
result = asc2.add(input, 2)

This function can also be called via a binary operator on LocalTensor, as input + other instead of add(input, other).