akhunti1[m] has quit [Remote host closed the connection]
<rcurtin[m]>
ok, everything is merged except #3491 for the next mlpack release; once that's ready I'll use the release script and maybe it will even work correctly :)
<zoq[m]>
<rcurtin[m]> "ok, everything is merged except..." <- Did we change something on the test suite? For https://github.com/mlpack/mlpack/pull/3491 I don't see which test case fails.
<rcurtin[m]>
I think that is just unlucky. Maybe it would be good to increase the size of the random matrix in that test, but I'm not sure if that will help or hurt or what
<zoq[m]>
That's what I thought, that it's a random test failure.
<rcurtin[m]>
I see you approved the PR just now, so whenever you merge that I'll run the release script for 4.2.0 (or you can do that if you like, it should work the same regardless of who does it)
<rcurtin[m]>
last release mlpack-bot just completely didn't do its job, so I think the Github API has probably changed or something, so I'll try and update the JS code before the merge happens
<zoq[m]>
Sounds good.
<rcurtin[m]>
I'm very close with Bandicoot, by the way... cross() and conv() are all that remain, and cross() is trivial (I'm not sure why I included it, it's actually not very suitable for GPUs, but I'll still do it because I'd like Bandicoot to have the complete Armadillo API eventually)
<rcurtin[m]>
conv() will probably be hard though... we'll see
<zoq[m]>
<rcurtin[m]> "conv() will probably be hard..." <- Really nice, I guess, time to update the open ensmallen PR.
<rcurtin[m]>
yeah, definitely. I think I've implemented most everything ensmallen should need, but I may have missed some things here or there