Stefan Kroboth

> However, on second thought, OWL-QN only adds a few processing to L-BFGS, and a separate structure would incur a high administrative cost since changes to L-BFGS would have to...

Such a benchmark would be really interesting. I haven't compared against these implementations at all, so besides the performance differences, it may be possible that they do not produce the...

This is a lot slower than I expected. > code optimization: as far as I understood form @stefan-k comment above, no one has spent much too time optimizing LBFGS in...

Thanks for opening this issue! This would be an ideal initial project for someone interested in contributing. I would of course provide mentoring/support if needed.

Great! :) Feel free to open a PR when you're ready or when there is something to discuss.

Hi, Thanks for your interest in contributing to argmin! CMA-ES would be a great addition to argmin and I'd very much appreciate it if you would implement this. We've recently...

Hi @Armavica, Have you had a chance to give this a try yet? Is there anything I can help you with?

Sounds great! :) There's no rush, take your time!

Hi @PyPete , thanks for your interest in contributing! I haven't heard anything from @Armavica in a while, so I'm not sure what the current status of this is. I'm...

@VolodymyrOrlov sounds great! :) The only other population based method so far is particle swarm optimization which uses `PopulationState` to carry state from one iteration to the next. `PopulationState` is...