Closed ylabbe closed 1 year ago
@ylabbe, some np.float
occurances are replaced by np.float32
and some by np.float64
. According to the documentation you linked, np.float
is equivalent to np.float64
. I therefore propose to replace all np.float
by np.float64
, to make sure all intermediate results stay identical.
Agree, I changed all occurrences of np.float to np.float64.
np.float is deprecated since numpy 1.20 https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations