shukon

On first glance this is likely unrelated to the `Warning` you got. Instead, please try running fANOVA with swig 3.0 - fANOVA builds on `pyrfr`, have a look at https://github.com/automl/random_forest_run/issues/51...

Save/Load random forest to avoid (accidental) tempering - binary - check pyrfr test code for example - timing!

As an update: While it's possible to save and load pyrfr's, it seems non-trivial to save and load the SMAC-RandomForests...

Looks good to me from a quick glance. Since it's only jupyter-methods I'd deem it safe to merge, even though travis blocks it... that's some import-struggle with smac methods, I...

Yeah... well I can actually just delete them from dev and see what happens ;)

Example implementation in https://github.com/automl/CAVE/tree/feature/cfp_parallel, however there seems be no speedup (actually it's slower by a factor of ~1.3) for the distances and a `MemoryError` occurs for the MDS...

Thanks a lot for bringing this up and providing an excellent MWE. This is similar to #96, it is indeed a bug and will now be fixed with #104.

Fixed in new release 2.0.19.

I still don't really get why we redirect output to `/dev/null`... I'm having trouble to debug issues with my `borg_ynh` - can you explain why we're discarding the output?

I did and there are logs in there... however I don't see any actual output on what *actually* fails, I don't even get the **Backup failed miserably** error message, however...