Incompatible with numpy >= 1.20 due to use of type aliases.

mdbloice / Augmentor

Image augmentation library in Python for machine learning.

https://augmentor.readthedocs.io/en/stable

MIT License

5.08k stars 866 forks source link

Incompatible with numpy >= 1.20 due to use of type aliases. #250

Closed jsilter closed 1 year ago

jsilter commented 1 year ago

See https://numpy.org/doc/stable/release/1.20.0-notes.html#using-the-aliases-of-builtin-types-like-np-int-is-deprecated

Type aliases only used in two places (that I could find).

mdbloice commented 1 year ago

Oh yes, you're right. Thanks for pointing this out.

I also noticed that I used np.uint8 a fair bit here and there (mostly in the tests), but it seems that things like np.uint8 and np.float16 etc. will not deprecated in order to allow for control of precision. So those can be kept. The two you found seem to be the only relevant ones that would need to be changed.