ChanServ changed the topic of #mlpack to: Due to ongoing spam on freenode, we've muted unregistered users. See http://www.mlpack.org/ircspam.txt for more information, or also you could join #mlpack-temp and chat there.
anckygP has joined #mlpack
anckygP has quit [Remote host closed the connection]
< mentekid>
Haven't used IRC in a while! What's this about freenode spam? :O
< rcurtin>
mentekid: yeah, there's been a long-running spam attack
< rcurtin>
I idle in the scikit-learn channel to keep my eye on it; they haven't taken any measures to try and clean it up
< rcurtin>
it looks like for now, the spambots are still just parroting random logs from other IRC channels, but it looks like the frequency is slowing down
< rcurtin>
by the way, saw the update to ensmallen#4, I don't think anyone else is working on it so feel free if you like :)
< mentekid>
Sounds pretty annoying. It's definitely better to get people to register their username and be free I think
< rcurtin>
yeah. and luckily the spambots are not very sophisticated, so we can set rules so that anyone connected via webchat can still talk even if not registerd
< rcurtin>
registered*
< rcurtin>
how is Cambridge? :)
< mentekid>
I'm trying to get the tests to compile, as I said I haven't used mlpack in a while
< rcurtin>
the mlpack tests, or ensmallen tests?
< mentekid>
Cambridge is nice, autumn is always a nice season here. Very cold though, it was 6 Celsius today
< rcurtin>
wow... that sure would be nice... here it's 31C right now...
< mentekid>
You're still in Georgia right?
< rcurtin>
yeah
< rcurtin>
I would have expected the summer to be over by now, but I guess not yet :(
< mentekid>
Summer lasting too long has never been a problem in Britain I think. Though this summer it was 35 for a couple of days and people were complaining.
< mentekid>
To be fair houses are not built for that heat, almost nobody has A/C
< rcurtin>
yeah, 35C is awful if there's no A/C... I think the entire culture of the southeast is built around A/C everywhere
< rcurtin>
I'm a happy participant of that :)
< mentekid>
Yeah I guess it's justified there - here not so much, and it's OK except for when it really really is not
< mentekid>
But it's fun to see the Brits try to cope with a month of no rain
< rcurtin>
ha
< rcurtin>
"what is this bright orb in the sky??"
< mentekid>
"And why is it burning?"
< mentekid>
For the tests - should I be looking at the tests/ or the tests_catch/ directory? The second one looks like work in progress
< mentekid>
(Judging from the makefile)
ImQ009 has quit [Ping timeout: 245 seconds]
< rcurtin>
yeah, I'm not sure---I think Conrad is working in the tests_catch/ directory, but I think Marcus and I were working in tests/
< rcurtin>
realistically though, I feel like if you make changes to tests/, it's easy enough to just copy the work over if Conrad wants to
< rcurtin>
also it'll be hard to test the changes for now, since a lot of the tests still don't compile
< mentekid>
Oh right, I didn't realize that
< mentekid>
I thought I was doing something wrong
< rcurtin>
yeah, mostly the tests have just been pulled out of mlpack and only modified slightly, which isn't enough yet :)
< mentekid>
So the link error for arma::arma_rng_cxx11_instance is expected?
< rcurtin>
yeah
< mentekid>
Ok, good to know
< rcurtin>
maybe linking with -larmadillo will fix it? I'm not sure
< rcurtin>
but I have yet to successfully run any of the tests
< rcurtin>
(they should all pass of course as soon as they're hooked up right, so I'm not worried)
< mentekid>
Yeah a few of them look like they still include a bunch of mlpack headers, so makes sense that they don't build
< rcurtin>
yep :)
< zoq>
For testing I use the converted sgd_test.cpp file, include #include "ensmallen_bits/problems/problems.hpp" in ensmallen.hpp and add the namespace ens::test.