pub trait Float: FromPrimitive + Float + PartialOrd + Sync + Send + Default + Display + Debug + Signed + Sum + NumAssignOps + AsPrimitive<usize> + for<'a> AddAssign<&'a Self> + for<'a> MulAssign<&'a Self> + for<'a> SubAssign<&'a Self> + for<'a> DivAssign<&'a Self> + MulAdd<Output = Self> + SampleUniform + ScalarOperand + AbsDiffEq {
type Lapack: Float;
fn cast<T: NumCast>(x: T) -> Self { ... }
}
Expand description
Floating point numbers
This trait bound multiplexes to the most common assumption of floating point number and implement them for 32bit and 64bit floating points. They are used in records of a dataset and, for regression task, in the targets as well.