On Mon, Apr 22, 2019 at 2:20 PM Ralf Gommers wrote:
>
>
> On Mon, Apr 22, 2019 at 9:26 PM Nathaniel Smith wrote:
>
>> Your last email didn't really clarify anything for me. I get that
>> np.func.__numpy_implementation__ is intended to have the semantics of
>> numpy's implementation of func, but
On Mon, Apr 22, 2019 at 9:26 PM Nathaniel Smith wrote:
> Your last email didn't really clarify anything for me. I get that
> np.func.__numpy_implementation__ is intended to have the semantics of
> numpy's implementation of func, but that doesn't tell me much :-). And
> also, that's exactly the de
Your last email didn't really clarify anything for me. I get that
np.func.__numpy_implementation__ is intended to have the semantics of
numpy's implementation of func, but that doesn't tell me much :-). And
also, that's exactly the definition of np.func, isn't it?
You're talking about ~doubling th
Are there still concerns here? If not, I would love to move ahead with
these changes so we can get this into NumPy 1.17.
On Tue, Apr 16, 2019 at 10:23 AM Stephan Hoyer wrote:
> __numpy_implementation__ is indeed simply a slot for third-parties to
> access NumPy's implementation. It should be con
On Thu, Apr 18, 2019 at 10:52 AM Stuart Reynolds
wrote:
> Is float8 a thing?
>
no, but np.float16 is -- so at least only twice as much memory as youo need
:-)
array([ nan, inf, -inf], dtype=float16)
I think masked arrays are going to be just as much, as they need to carry
the mask.
-CHB
>