Fix np.float_ to np.float64 while using numpy >= 2.0

[geoochi]

Overview: What does this pull request change?

use np.float64 while numpy >= 2.0

Motivation and Explanation: Why and how do your changes improve the library?

Links to added or changed documentation pages

Further Information and Comments

Reviewer Checklist

• The PR title is descriptive enough for the changelog, and the PR is labeled correctly
• If applicable: newly added non-private functions and classes have a docstring including a short summary and a PARAMETERS section
• If applicable: newly added functions and classes are tested

[JasonGrace2282]

Is there a reason not to use np.float_ in numpy>=2.0?

[geoochi]

reason: https://numpy.org/devdocs/release/2.0.0-notes.html

actually it will raise AttributeError when using np.float_ in numpy>=2.0:

AttributeError: np.float_ was removed in the NumPy 2.0 release. Use np.float64 instead.

[JasonGrace2282]

Oh I see, I missed it before. Thanks :)