naywhayare changed the topic of #mlpack to: http://www.mlpack.org/ -- We don't respond instantly... but we will respond. Give it a few minutes. Or hours. -- Channel logs: http://www.mlpack.org/irc/
govg has quit [Quit: leaving]
jenkins-mlpack has joined #mlpack
udits has joined #mlpack
govg has joined #mlpack
< naywhayare>
udits: hello there! I haven't had a chance to respond to your email yet
< naywhayare>
I've spent the past couple days moving my desk and the whole mlpack build system, and I'm finishing getting it set up now
< naywhayare>
how are things going?
< udits>
hi ryan !
< udits>
wokay ! no problem... things have settled down now.
< udits>
even I reset my system, had to set up the entire mlpack build once more.
< naywhayare>
hopefully it wasn't that hard. :) if so, maybe we need to make some changes to simplify the build process
< udits>
how are things at your end ? Been travelling lately ? This semester must have started.
< udits>
Not at all...
< naywhayare>
no traveling for me, just work and more work :(
< naywhayare>
but, I do enjoy my work, so it's not that bad :)
< udits>
:)
< naywhayare>
I noticed today that there are lots of warnings in the adaboost and decision stump code:
< naywhayare>
it's mostly really simple stuff -- comparison between signed and unsigned
< naywhayare>
so for the most part fixing these entails changing int to size_t
< naywhayare>
anyway, I was curious if you were interested in fixing them
< naywhayare>
if not, I'll get around to it eventually... :)
< udits>
Ah, sorry. Sure, I will. Although I'm surprised they weren't thrown while building.
< naywhayare>
it depends on the compiler and versions of various libraries
< naywhayare>
so it may be that those warnings don't appear in your setup, but they do on the build server
< udits>
Hmm. I shall get to that then.
< udits>
So, I was wondering. I have been reading up on Deep Learning and Convo Neural Nets, somewhat... do you guys plan to implement/extend towards that area ?
< naywhayare>
Siddharth started implementing something at the beginning of the summer
< udits>
I was going through theano and deeplearning4j. I'm interested enough.
< naywhayare>
he wrote a sparse autoencoder
< udits>
Go on...
< naywhayare>
he's also written softmax regression, but it has a bug that we figured out a few weeks ago
< naywhayare>
but I haven't seen him in a little while, and I don't think he's committed the fix yet
< naywhayare>
he's probably busy with his job... I think he is working for Amazon in Bangalore right now (I'm not certain)
< naywhayare>
anyway, I'm not sure what his plans after that were
< udits>
Hmm... okay. I'll see what I can do, come up with something I want learn and implement as well...
< naywhayare>
you could probably send him an email or find him on IRC one of these days to figure out what his plan is
< udits>
yeah.
< jenkins-mlpack>
Starting build #2114 for job mlpack - svn checkin test (previous build: STILL UNSTABLE -- last SUCCESS #2109 15 days ago)
udits has quit [Quit: Leaving]
< marcus_zoq>
udits: Looks like now that the ImageNet challenge is over more and more people are interested in convents, have you seen the prediction of the google submission: http://cs.stanford.edu/people/karpathy/ilsvrc/
< marcus_zoq>
ah too late ...
< naywhayare>
ah well, I'm sure he'll be back soon enough
< marcus_zoq>
If you are interested I can add the neural networks that I've written for my thesis.