Optimizing a Gaussian penalty function for HSL color compatibility in PyTorch/NumPy

This seems to be a question about optimizing some code to increase its speed.

The code you have posted looks trivial. Even in python, it should run almost instantaneously

So:

• tell us how fast it runs on some non-trivial input
• tell us how much faster you need the code to run.

I don't know about the details of this color harmony system, but the usual way to compare angular values is with trigonometric functions. sin(0) = 0, sin(90) = 1, cos(0) = 1, etc... TensorFlow probably has an efficient implementation of those. On the other hand, the implementation you have may well be more efficient, you would probably have to benchmark your code to be sure.

I don't know about the details of this color harmony system, but the usual way to compare angular values is with trigonometric functions. sin(0) = 0, sin(90) = 1, cos(0) = 1, etc... TensorFlow probably has an efficient implementation of those. On the other hand, the implementation you have may well be more efficient, you would probably have to benchmark your code to be sure.

np.exp(- (hue_diff 2) / (2 sigma 2))

Here, 2 sigma 2 is a constant for every array value. Is you compiler clever enough to optimize this away? Why not pass it in as a parameter, rather than just sigma?

I don't know about the details of this color harmony system, but the usual way to compare angular values is with trigonometric functions. sin(0) = 0, sin(90) = 1, cos(0) = 1, etc... TensorFlow probably has an efficient implementation of those. On the other hand, the implementation you have may well be more efficient, you would probably have to benchmark them both to know which is better.