ParBayesianOptimization

ParBayesianOptimization copied to clipboard

Fix #52: removing `setorder` due to reproducibility-issues

6

This PR fixes #52 by removing `setorder` which is currently applied to the `LocalOptims` object and leads to inconsistent results on different hosts when ties occur in the sorting criteria...

kapsner

Error in rbindlist(scoreSummary) : Column 2 of item 1 is length 3 inconsistent with column 1 which is length 100. Only length-1 columns are recycled.

1

I get this error right after running initial score function when initPoints is greater than a 100. > Error in rbindlist(scoreSummary) : > Column 2 of item 1 is length...

marjohnsen

Reproducibility with parallelization

8

Hi, first of all thanks for that great package, which speeds up hyperparameter optimization enormously!! I am currently stuck with getting reproducible results for xgboost on different machines using parallelization....

kapsner

clustermq integration?

1

The README shows how to parallel on a single machine via doParallel, but it would be great if the docs included info on a multi-node HPC run of ParBayesianOptimization via...

nick-youngblut

May I inquire about the possibility of writing a **user input acquisition function** and passing it into the function calcAcq()? For instance, if a user defines beta()

xinyuz1996

Multi Objective Optimization

1

Hello! Using the "ParBayesianOptimization" package, it possible to use this package for "multi objective optimization" (e.g. optimize several cost functions together)? For example, below I have included an example of...

swaheera

Any way to add constraints in between the parameters?

3

Hi, First off, thanks for your great package! I have used it to tweak hyperparameters. I thought about your package to solve an ensembling optimization issue. I was wondering if...

actuarial-lonewolf

**Problem** It seems that currently, I can't provide an `initGrid` consisting of a data frame with only 1 row. The `inBounds

juliuskittler

add na.rm=TRUE to min/max in zeroOneScale

1

Thanks for developing this, it's a cool package! I came across a small bug where if the user-provided `FUN` returns `list(Score=NA)` during fitting, I get the following error: ``` Error...

kevingoneill

Parallel Acquisition functions

5

Hi could you please look into adding parallel acquisition functions such as q-EI or q-UCB? Thanks for your great work so far!

feji3769