tf.math.nextafter

Returns the next representable value of x1 in the direction of x2, element-wise.

tf.math.nextafter(
    x1, x2, name=None
)

This operation returns the same result as the C++ std::nextafter function.

It can also return a subnormal number.

x1 A Tensor. Must be one of the following types: float64, float32.
x2 A Tensor. Must have the same type as x1.
name A name for the operation (optional).

A Tensor. Has the same type as x1.

Cpp Compatibility

Equivalent to C++ std::nextafter function.