remainder_op

→TensorImplPtr

int remainder_op(const int & a, const int & b)

Python-style modulo: $y = a - \lfloor a/b \rfloor \cdot b$ .

Composite over div_op + floor_op + mul_op + sub_op. Result has the sign of b because floor rounds toward $-\infty$ .

Math

y = a - \lfloor a/b \rfloor \cdot b

See Also

fmod_op — C-style sign convention.

a, bTensorImplPtr

Broadcastable operands. b must be non-zero where evaluated.

TensorImplPtr

Tensor on the broadcast shape with the sign of b.

Loading…