View source on GitHub
|
Constructs a complex tensor whose elements are Cartesian coordinates corresponding to the polar coordinates with absolute value abs and angle angle.
tf.keras.ops.polar(
abs_, angle
)
The operation is numerically equivalent to torch.polar().
It is not equivalent to scipy.lingalg.polar() which performs
Singular Value Decomposition.
Given the magnitude (abs_) and angle (angle), this function computes the
corresponding complex number in the form of real + imaginary * 1j, where:
real = abs_ * cos(angle)imaginary = abs_ * sin(angle)
Args | |
|---|---|
abs_
|
The magnitude (absolute value) of the complex number. |
angle
|
The angle (in radians) of the complex number. |
Returns | |
|---|---|
A complex number (or array of complex numbers) with the same shape as
abs_ and angle.
|
Example:
abs_ = keras.random.normal((1, 2))angle = keras.random.normal((1, 2))keras.ops.nn.polar(abs_, angle).shape(1, 2)keras.ops.nn.polar(abs_, angle)Array([[0.63185346-0.59370506j, 0.48960376-0.31677645j]], dtype=complex64)