Erik Welch

I agree that hyper-parameter searches are an important use case (dask/dask-searchcv#32 actually prompted me to finally start this project) that we test and show off. > For general problems you...

Here are the SciPy benchmark functions: https://github.com/scipy/scipy/tree/master/benchmarks/benchmarks/go_benchmark_functions

This may have some useful machine learning hyper-parameter benchmarks: http://automl.chalearn.org/

I would also like for different optimization runs to be able to share the same underlying grid and results cache. This would get us one step closer to creating a...

I think our best option is to allow categoricals to be explored and optimized *outside* of our main `search` function, which will require cooperative multitasking (#15). In other words, categorical...

I just love those performance reports! This does look suspicious--thanks for the report. I don't yet have a plausible explanation for why some threads would stall. The simplest "I have...

I can reproduce the issue. I have ruled out some of the weird bits of `afar`, such as using locks around updating `builtins.print` and using a custom channel to send...

I agree 100%. That would be great, and looks like it's doable. Related: #410.

Also, I really like small, achievable suggestions like this. We all know documentation needs improved, but it often seems like a large, daunting task. Please don't be shy about more...

Yeah, this is worth attention, thanks @szhorvat. Right now we defer to `networkx.to_scipy_sparse`, but they "solved" this problem by adding to the documentation and providing a recipe for how to...