As of numpy==2.0, the cython np.int_t type no longer exists. Instead, an explicitly-sized type such as np.int32_t or np.int64_t must be used. Here's a relevant stack overflow post:
lie_learn currently uses np.int_t in one place, and as a result cannot be installed with up-to-date numpy. The PR replaces this one usage with np.int64_t, which should mean the exact same thing.
As of
numpy==2.0
, the cythonnp.int_t
type no longer exists. Instead, an explicitly-sized type such asnp.int32_t
ornp.int64_t
must be used. Here's a relevant stack overflow post:https://stackoverflow.com/questions/21851985/difference-between-np-int-np-int-int-and-np-int-t-in-cython
lie_learn
currently usesnp.int_t
in one place, and as a result cannot be installed with up-to-date numpy. The PR replaces this one usage withnp.int64_t
, which should mean the exact same thing.