expand_op

→TensorImplPtr

int expand_op(const int & a, const int & shape)

Alias for broadcast_to_op with reference framework's expand semantics.

Provided so callers porting from reference framework can keep the call site unchanged. Behaviour is identical to broadcast_to_op.

See Also

broadcast_to_op.

aTensorImplPtr

Input tensor.

shapeShape

Target shape (NumPy-broadcast-compatible with a.shape).

TensorImplPtr

Broadcast tensor.