Do not have explicit dtype in hints

[juanitorduz]

When trying to fix #210 I am seeing a lot of casting code which can be distracting and really not important. What if we use something like

https://numpy.org/doc/stable/reference/typing.html#number-precision

or even

from jaxtyping import Array

from https://docs.kidger.site/jaxtyping/ (despite the name of the package we can just use it for NumPy)

JAX has a nice array convention, but I do not see this in NumPy (ArrayLike, yes but not purely Array), see https://jax.readthedocs.io/en/latest/jep/12049-type-annotations.html

Thoughts?

[aloctavodia]

not sure which one is a better solution. But I agree with your diagnostics.

[juanitorduz]

Ok! Let me do some research these days to have better data and options 🤔

[juanitorduz]

Actually, in dd6a6e8 I just removed the reference to the specific float precision (np.float64) and things look cleaner.