Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-02-21

| Channels | #pdl index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
04:40 meagan joined #pdl
04:40 mad_hatter hello friends!
04:53 sivoais hi mad_hatter!
04:53 mohawk hello
04:53 mohawk what's up?
04:54 mad_hatter not much
04:55 mad_hatter I was just curious about a perl module that provided methods to accomodate matrix algebra and was pointed here
04:55 mohawk ok
04:55 mohawk pdl is certainly the sharpest knife there is on that one
04:56 mohawk ...?
04:57 sivoais yeah, PDL is just one the ones out there. It is really good at performing quickly on large amounts of data.
04:59 mohawk mad_hatter, do you have any questions or issues on pdl, or are you just basking in the glory of it all?
05:02 sivoais hehe, if you have any questions, just ask away
05:02 mad_hatter neat
05:02 mohawk i have a question
05:02 mad_hatter well...im trying to build a pretty neat neural network
05:02 mohawk pdl 2.007 was like released ages ago
05:02 mohawk is a new version ever going to be released?
05:03 mohawk go on
05:04 mad_hatter and I'm trying to determine the correct way to go about weighting inputs
05:06 sivoais are you learning the weights?
05:07 mad_hatter yes
05:07 sivoais backpropagation?
05:08 mohawk i've got nothing to add to that, good luck - sivoais is truly brilliant and if anyone can help, he can
05:09 mad_hatter sivoais, yes
05:14 sivoais mad_hatter: would you be able to using something like <http://leenissen.dk/fann/wp/>? If you are doing this for learning, sure program it in Perl, but otherwise already written code is already tested. :-)
05:18 mad_hatter well i would
05:18 mad_hatter but I don't know C
05:18 mad_hatter and I know theres a module AI::Fann for Perl but I don't think it's very robust
05:18 mad_hatter but this is more about learning
05:18 mad_hatter than using it in a productive way
05:19 sivoais Ah, I see. Makes sense then! Let me see if I can find some examples. I haven't done much NNs myself. I mostly reach for SVMs when doing classification/regression.
05:20 * sivoais makes a mental note to see about improving AI::FANN later
05:53 sivoais mad_hatter: looking around, maybe these might help <http://www.cse.unsw.edu.au/~cs9417ml/MLP2/>, <http://neuralnetworksanddeeplearning.com/chap2.html>
06:00 sivoais I'm a little busy right now, but if you share your code, I can try taking a look in a couple days
06:17 mad_hatter sure thing sivoais
14:55 mad_hatter joined #pdl
17:04 mad_hatter joined #pdl
17:05 mohawk i remember doing NNs in one of my compsci modules back in 93
17:06 mohawk which may just be before sivoais was born? :-)
17:37 sivoais mohawk: haha, not quite, but very close
17:43 jberger oh oh, I was 10, I could have done it :P
18:04 mad_hatter joined #pdl
19:05 mohawk ha ha
19:05 mohawk i'm zapping mymalloc.p, it's pointless
19:06 mohawk just put the relevant code in pdl.h.PL
19:07 mohawk that's also going to just become a .h, with some relevant -D action
19:11 mohawk next up, pdl_hfiles.p
19:11 mohawk the functionality should be in Types.pm as a method to return relevant header text
19:34 mohawk next target: badsupport.p
19:38 mohawk that info is all in PDL::Config
19:53 mad_hatter joined #pdl
19:58 mohawk well that was surprisingly easy
22:13 opkick [pdl] wchristian pushed 4 new commits to master: http://git.io/AEmM
22:13 opkick pdl/master 1ca5d22 Petr Písař: Do not install modules using PDL::Slatec if Slatec is disabled...
22:13 opkick pdl/master 30a6cd4 Ed J: Remove no-longer-.PL from MANIFEST.SKIP
22:13 opkick pdl/master bc1cf2d Ed J: Fix Inline::Pdlpp for Cwd on Windows "quirk"
22:26 jberger mohawk++
22:34 * sivoais goes to rebase the Travis-CI code
22:40 sivoais hah, I was just about to e-mail the list about merging the Travis-CI code, but chm got to it first :-D
22:43 opkick [pdl] wchristian pushed 17 new commits to master: http://git.io/AECV
22:43 opkick pdl/master 3d9fc29 Zakariyya Mughal: add .travis.yml
22:43 opkick pdl/master 8b99eea Zakariyya Mughal: add native dependencies to .travis.yml
22:43 opkick pdl/master ac13231 Zakariyya Mughal: make code coverage run under Perl v5.20

| Channels | #pdl index | Today | | Search | Google Search | Plain-Text | summary