Function dfdx::tensor_ops::nans_to

source ·

pub fn nans_to<S: Shape, E: Dtype, D: UnaryKernel<NansToKernelOp<E>, E>, T: Tape<E, D>>(
    t: Tensor<S, E, D, T>,
    value: impl Into<f64>
) -> Tensor<S, E, D, T>

Expand description

Replaces any std::f32::NAN with value.

Pytorch equivalent: t.nan_to_num(value)

Example:

let t = dev.tensor([1.0, f32::NAN, f32::NAN, 4.0]);
let r = t.nans_to(0.0);
assert_eq!(r.array(), [1.0, 0.0, 0.0, 4.0]);