Closed thelmuth closed 4 years ago
Maybe this is also the time to remove some gigantic classification data sets from Penn... Is anyone working with these? I am not.
There's one file, a_5000s_2000her_0.1__maf_0.2_EDM-1_01.txt
, which is 20MB (i.e. most of the Clojush directory's space besides .git
). The other four files there are all just 68KB each, so they aren't breaking the bank. I'm fine either way!
BTW, I'm working on another pull request that implements batch lexicase selection, and I noticed that one of your most recent pushes makes it print "Preserving frontier... " every generation, even when that setting isn't used. I assume it's ok to move that to within the relevant when
to clean it up?
Yes, all of this sounds great. Push a commit with all of these changes and I'll merge it.
Ok, will do in a minute. It will be a new Pull request; closing this one.
… better logging of code.
I'm not sure if anyone but me is using fly_launcher these days, but if so, this should be useful for them.
The current Clojush directory is about 100MB, about 3/4 of which is taken up by .git. We could leave .git off, but I don't think the 75MB per set of runs is going to make much difference. (If I had heard that as a kid when we had a "big" 40GB hard drive, I would have thought the current me were nuts.)