verne.freenode.net 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/
devang has joined #mlpack
devang has quit [Client Quit]
travis-ci has joined #mlpack
< travis-ci>
mlpack/mlpack#591 (master - a69871c : Ryan Curtin): The build passed.
< kirizaki>
always You can use mlpack mailing list and ask here
< kirizaki>
for sure someone will help You :)
< Guest19631>
thanks for the prompt response kirizaki :)
< rebeka>
Hello. I am Rebeka Mukherjee from India. I am interested to take part in GSoC 2016. Some of the topics I have studied previously and am intersted to work on are k-center algorithm, k-nn algorithm, neighbourhood-based collaborative filtering, decision trees, deep learning modules. Could someone please get back to me with more information on these projects so that I can decide my topic for the proposal?
< kirizaki>
no problem, I'll be out right now 'cause I'm sick but: We don't respond instantly... but we will respond. Give it a few minutes. Or hours. -- Channel logs: http://...
< kirizaki>
and that's all for now with which I can help You, because I'm not the one who knows a lot about those algorithms
< rebeka>
Thank you. I have gone through the github link for GSoC. I will now delve into the tutorials :)
< kirizaki>
but Your question will be responded because "rcurtin" is the main guru xD
< kirizaki>
go ahead ;)
praveench has joined #mlpack
wasiq has joined #mlpack
tsathoggua has quit [Quit: Konversation terminated!]
rohitpatwa has joined #mlpack
Guest19631 has quit [Quit: Page closed]
mentekid has joined #mlpack
cache-nez has joined #mlpack
rohitpatwa has quit [Quit: No Ping reply in 180 seconds.]
rohitpatwa has joined #mlpack
devang has joined #mlpack
praveench has quit [Ping timeout: 252 seconds]
rohitpatwa has quit [Ping timeout: 264 seconds]
vineet has joined #mlpack
devang has quit [Quit: Page closed]
Nilabhra has joined #mlpack
< Nilabhra>
anyone here?
vineet has quit [Ping timeout: 248 seconds]
< mentekid>
I am here but I am a GSoC student not a developer :P
hunter__ has joined #mlpack
< Nilabhra>
mentekid: hi, have you chosen a project?
anveshi has joined #mlpack
anveshi_ has joined #mlpack
< mentekid>
Nilabhra: I am thinking about the Approximate Nearest Neighbors one, I've worked with such algorithms before. How about you?
< Nilabhra>
mentekid: thinking about working on collaborative filtering
< Nilabhra>
mentekid: do you know how many tickets did mlpack get for 2016?
< mentekid>
Nilabhra: by tickets you mean applications? I have no idea
< Nilabhra>
mentekid: okay :)
< mentekid>
Nilabhra: the ML seems quite active though, so if everyone sending emails is applying I guess they will get a lot of them
rebeka has quit [Ping timeout: 252 seconds]
< mentekid>
have you compiled the system? I am looking around for documentation and stuff, I've run a few algorithms but haven't found a way to properly run test code
< Nilabhra>
no google allows organizations to let only a few students to work on gsoc. I was asking if anyone knows the number of students google will allow for mlpack.
< Nilabhra>
No I just saw mlpack today, will download and compile soon enough.
< Nilabhra>
mentekid: ^
< mentekid>
Nilabhra: Ah, I didn't know that, this is my first time in the programme. I assumed that #ideas = #students, probably I was wrong. I should read the programme information more carefully :P
< Nilabhra>
mentekid: :P yep totally
Nilabhra has quit [Remote host closed the connection]
rohitpatwa has joined #mlpack
LimeTheCoder has joined #mlpack
tejal567 has joined #mlpack
vasanth has joined #mlpack
tejal567 has quit [Quit: Page closed]
hunter__ has quit [Ping timeout: 252 seconds]
mentekid has quit [Ping timeout: 244 seconds]
vasanth has quit [Quit: Page closed]
< kirizaki>
mentekid: "but haven't found a way to properly run test code" if You mean to run mlpack test
< kirizaki>
You shoud do: ./build/bin/mlpack_test -p
agobin_ has joined #mlpack
anveshi has quit [Remote host closed the connection]
anveshi_ has quit [Remote host closed the connection]
ivy has joined #mlpack
ivy has quit [Ping timeout: 252 seconds]
Nilabhra has joined #mlpack
anveshi has joined #mlpack
< Nilabhra>
kirizaki: so how to get started? Should I first do some bug fixing?
< kirizaki>
You can familiarize Your self with Issue list at github ;)
agobin_ has quit [Quit: Connection closed for inactivity]
na1taneja2821 has joined #mlpack
tafodinho has joined #mlpack
< tafodinho>
hello everyone am a computer engineering student who wish to participate in the 2016 GSoC program with mlpack as my mentoring organization please hw do i get started
< rcurtin>
oops, that link to the preselected one should be http not https
< rcurtin>
thanks :)
< rcurtin>
I think that all the benchmarks need to be run again for new versions of mlpack and other libraries, but our build server setup is in a state of transition right now
< rcurtin>
I'm looking on our build system, I think that the benchmarks you're looking at were done with shogun 3.2.0
< rcurtin>
I can see that the command used was this:
< rcurtin>
make CONFIG=config.yaml BLOCK=shogun LOG=True run PYTHONPATH=/opt/scikit-learn/scikit-learn-0.13.1/lib/python3.3/site-packages/:/opt/mlpy/mlpy-3.5.0/lib/python3.4/site-packages/:/usr/local/lib/python3.4/dist-packages/ MS_PRINT_BIN=/usr/bin/ms_print VALGRIND_BIN=/usr/bin/valgrind LD_LIBRARY_PATH=/usr/local/lib/ LIBRARY_PATH=/usr/local/lib/ UPDATE=False
< rcurtin>
but that only specifies versions of other libraries and doesn't seem to specify the shogun version... so my guess is that the config.yaml you linked to before is the version that was used
< wiking>
kk
< wiking>
so yeah we need updates :)
< wiking>
we should create a dedicated site for this
< wiking>
benchmarks.ml
< wiking>
:)
< wiking>
and generate there the outputs
< zoq>
wiking: yeah, right, we use massif to run a memory benchmark for mlpack ... but it should be straightforward to run memory benchmarks for the rest of the libs ... I'm here to help you out :)
< rcurtin>
this would definitely be possible, and the existing benchmarking system would be good for this
< rcurtin>
I'm happy to supply dedicated benchmarking systems, but it'll be a little while until I can do that
< wiking>
that's alright
< wiking>
i'll put together couple of bots
< wiking>
and i'll ping if i'm stuck
< rcurtin>
one of the problems with using something like EC2 for benchmarking is that it'll be a VM running on a system that's running lots of other VMs, so depending on the load of the whole system, you may get noisy benchmarks
< wiking>
thanks so much for the help
< wiking>
aws is a shit :)
ankur has joined #mlpack
< rcurtin>
sure, I'm happy to help if you have any issues :)
< wiking>
i mean for benchmarking
< rcurtin>
yeah
< wiking>
those fake qemu things
< zoq>
:)
< wiking>
sometimes dont even have the power they say they have
< wiking>
did some benchmarks
< rcurtin>
it's really nice because it's free for open-source projects (at least some amount is), but you can't really use it for benchmarking or timing
< wiking>
actually rackspace is giving you much better vps
< wiking>
free?
< wiking>
aws is free for open source?: )
< wiking>
yeah we have some dedicated machines
< wiking>
found that to be the most cost efficienct solution
< rcurtin>
I thought AWS was free for some open-source use, I forget
< rcurtin>
I don't remember, that was years ago I think
< rcurtin>
could be different now :)
< wiking>
ah dunno
< wiking>
never used it for such purposes
< wiking>
anyhow
< rcurtin>
neither did I, I think I just read it somewhere
< wiking>
let's see how we can get some nice updates on this :)
< rcurtin>
yeah, I have always wanted to put more time into this project but never had the time
< wiking>
hehehe
< rcurtin>
there's a lot of potential for a really slick setup :)
< zoq>
kirizaki: Nice to hear that you feeling better. Haven't had time to configure the system, probably tomorrow. Anyway, I tested the code on FreeBSD and it builds without any problems.
< zoq>
palashahuja: Hello, can you point me to the test you are using?
< palashahuja>
Sure, I will create the gist if you'd like
< rcurtin>
zoq: if it works on freebsd, feel free to go ahead and merge the PR (or let me know and I'll merge it)
< rcurtin>
ideal output should be that it compiles, and there's nothing about libbfd or libdl in the cmake output
anveshi has quit [Remote host closed the connection]
< zoq>
palashahuja: Are you sure this is the latest version? I can't compile the code.
< zoq>
rcurtin: okay, I'll check that later.
< palashahuja>
what's the error ?
mentekid has quit [Ping timeout: 244 seconds]
rohitpatwa has quit [Ping timeout: 240 seconds]
< zoq>
palashahuja: "no matching function for call to 'randu'" -> mask = arma::randu < arma::Cube < eT > > (baseLayer.Weights().n_rows, baseLayer.Weights().n_cols);
< zoq>
palashahuja: That's just one error. I can fix everything, but maybe I don't have to do that.
< palashahuja>
that should be Mat instead of cube
< palashahuja>
My bad ...
< palashahuja>
I'll update the code
< zoq>
palashahuja: Another error is: "dropconnect_layer.hpp:141:17: error: no matching member function for call to 'Backward'" -> baseLayer.Backward(gy, g);
< palashahuja>
I am sending the updated code ..
< zoq>
palashahuja: okay, great.
ranjan123 has quit [Ping timeout: 252 seconds]
chick_ has joined #mlpack
LimeTheCoder has quit [Ping timeout: 252 seconds]
mentekid has joined #mlpack
mentekid has quit [Remote host closed the connection]
travis-ci has joined #mlpack
< travis-ci>
mlpack/mlpack#597 (master - fd84f3a : marcus): The build was fixed.