Marius Lindauer

So, this concerns only using SMAC2? I would expect that in most cases 10 digits should be more than enough.

I think it would not hurt to fix this at some point, but it is not crucial.

on my machine I need roughly 5-10 iterations of the for loop s.t. the program gets stuck and requires tremendous amounts of memory (several tens of GB)

Hi Eddie, Could we provide these kind of additional information (inference time, model size, etc) simply as optional dictionary argument to the metrics? We also don't have to recompute this...

Hi, HpBandSter is the original implementation of BOHB. However, the HpBandSter repo is no longer actively maintained by us after the main developer left the group. Therefore, we reimplemented the...

@ffalfred Thanks for reporting this issue. @mfeurer since a user does not see the memory limit and runtime limit for each TAE call through the `fmin_smac` interface anyway, this sounds...

> What do you think? I thought you wanted the fmin-interface to be very similar to the scipy fmin!? But I'm fine with adding two more arguments (mem_limit, time_limit).

@mfeurer how to resolve this issue? We still use the pynisher in our fmin interface (if I'm not wrong). Should I simply disable it in the fmin interface?

Dear @KOLANICH , Thank you for this proposal. It is also on our agenda for quite some time, but we have not implemented it so far. Since this feature request...

@mfeurer could you please reply to it.