Why requires NumPy==1.11.0?

[aht]

Is there a specific problem with other versions? For example, our infrastructure uses numpy-1.13.3 and installing hyperopt-sklearn basically downgrades numpy.

It seems both sklearn and hyperopt only specifies generic numpy dependency.

[jaberg]

i suspect a packaging error, @brent?

…

On Wed, Oct 18, 2017 at 4:06 PM, Hai-Anh Trinh ***@***.***> wrote: Is there a specific problem with other versions? For example, our infrastructure uses numpy-1.13.3 and installing hyperopt-sklearn basically downgrades numpy. It seems both sklearn and hyperopt only specifies generic numpy dependency. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#86>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAKdDD_aNxyWDSvZl_LSN2EFVOOHdU69ks5stlpQgaJpZM4P-SH-> .

[gigglegrig]

same error with Numpy. It's uninstalling my Numpy1.13 and replace with Numpy 1.10, but failed.

[bjkomer]

There was an issue that came up with Numpy 1.12, so a workaround was to fix it to 1.11 where it worked.

This was a while ago so it's possible the newer numpy 1.13 is okay. If not, we'll have to figure out where it is coming from and fix it on our end.

[gigglegrig]

Yeah, I realized that too. The problem is Python 3.6 doesn't run with Numpy 1.11.
Hope this can be addressed~ Thanks

[aht]

@bjkomer I can verify that auto-sklearn works with numpy-1.13

[bjkomer]

Thanks! I've verified it as well, and hyperopt-sklearn works with numpy-1.13. The setup.py has been updated