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/
decltypeme has quit [Ping timeout: 260 seconds]
decltypeme has joined #mlpack
ank_95_ has joined #mlpack
hercky_ has joined #mlpack
hercky has quit [Ping timeout: 252 seconds]
hercky_ has quit [Ping timeout: 252 seconds]
drktrx_ has joined #mlpack
Jayaganesh has joined #mlpack
Jayaganesh has quit [Client Quit]
hunter has joined #mlpack
hunter is now known as Guest38900
Guest38900 has quit [Client Quit]
archit has joined #mlpack
< tsathoggua>
Hi, I am having problems building mlpack with OpenBLAS... and without installing anything in the system (that is, specifying local directories)
< tsathoggua>
I build armadillo directly invoking cmake with -DOpenBLAS_LIBRARY=blahblah/libopenblas.so
< tsathoggua>
for the mlpack cmake, I specify ARMADILLO_LIBRARY, ARMADILLO_INCLUDE_DIR (and also the BOOST locally compiled directory)
< tsathoggua>
I had no problems when I installed those libraries, but on a system with no root access, mlpack's cmake complains about the lack of LAPACK / etc
anveshi has joined #mlpack
< tsathoggua>
Ok... I had been looking at the wrong places all the time (mlpack was getting the defines from the wrong armadillo_bits/config.hpp). Sorry and thanks anyway!
drktrx_ has quit [Ping timeout: 252 seconds]
archit has quit [Ping timeout: 252 seconds]
sanket has quit [Ping timeout: 246 seconds]
sanket has joined #mlpack
mrbean has joined #mlpack
deepti has joined #mlpack
deepti has quit [Quit: Page closed]
Yen has joined #mlpack
< Yen>
Hello guys
sanket has quit [Ping timeout: 248 seconds]
tsathoggua has quit [Quit: Konversation terminated!]
decltypeme has quit [Quit: Leaving]
Yen has quit [Ping timeout: 252 seconds]
riddhish has joined #mlpack
riddhish has quit [Ping timeout: 252 seconds]
acbull_ has joined #mlpack
acbull_ has quit [Ping timeout: 252 seconds]
tsathoggua has joined #mlpack
anveshi_ has joined #mlpack
anveshi_ has quit [Client Quit]
anveshi_ has joined #mlpack
Guest92392 has joined #mlpack
Guest92392 has quit [Remote host closed the connection]
Guest92392 has joined #mlpack
anveshi has quit [Ping timeout: 252 seconds]
anveshi_ has quit [Ping timeout: 252 seconds]
mrbean has quit [Ping timeout: 246 seconds]
parijat has joined #mlpack
mrbean has joined #mlpack
tsathoggua has quit [Remote host closed the connection]
tsathoggua has joined #mlpack
tsathoggua has quit [Quit: Konversation terminated!]
parijat has quit [Ping timeout: 252 seconds]
sanket has joined #mlpack
parijat has joined #mlpack
< aditya123>
Do I start contributing using the mlpack version available for download or cloning the repository ?
< aditya123>
I was able to install mlpack from the tar version, but had issues with cmake in mlpack from repository
yvtheja has joined #mlpack
yvtheja_ has joined #mlpack
yvtheja_ has quit [Client Quit]
tdgunes has joined #mlpack
< yvtheja>
Hi everyone, could I please know the nick of Marcus Edel
viswanadh has joined #mlpack
ankur has joined #mlpack
Ashutosh has joined #mlpack
ankur has quit [Client Quit]
Manchv has joined #mlpack
< zoq>
yvtheja: That would be me :)
nikola-miljkovic has joined #mlpack
< yvtheja>
zoq, Hi, Really thanks for your reply :)
< zoq>
aditya123: If you are going to start contributing you should use the master branch. If, you tell use more about your problems, we could help solving the issue.
< zoq>
yvtheja: Great, I really like the project, in fact I'm about to setup a test system
< yvtheja>
zoq, Can I PM you ?
< zoq>
yvtheja: sure, but I'd like to keep the discussion as open as possible
Ashutosh has quit [Quit: Page closed]
< yvtheja>
zoq, Okay, I have good enough experience in ML using NN
< yvtheja>
zoq, but I am new to Deep Learning
< yvtheja>
zoq, I have just started taking Google Deep learning course
< yvtheja>
zoq, I want to if I can take up this project and learn things on the way.
< yvtheja>
zoq, I saw the project without Deep leaarning also but I am more interested in exploring Deep Learning
< zoq>
yvtheja: That's no problem at all, as pointed out in the wiki the "necessary knowledge" sections can be replaced with "willing to learn". So I guess a good start would be read some papers about the neuroevolution algorithms mentioned in the project idea. You should also try to build mlpack and the code to communicate with the nes emulator.
< zoq>
I'm about to write a readme for the nes emulation code which could be found here: https://github.com/zoq/nes. You can run the code on your machine or use our system.
< yvtheja>
zoq, Sure, I will run and understand the things
< yvtheja>
zoq, Thanks a lot for the info :)
< zoq>
yvtheja: Sure, if you have any further questions please don't hesitate to ask.
acbull has joined #mlpack
< aditya123>
zoq: I am getting the following error: Could not find the following Boost library : boost_serialization
anshumitts has joined #mlpack
< zoq>
aditya123: Can you check that you installed all dependencies https://github.com/mlpack/mlpack#3-dependencies especially boost-serialization. On debian the package should be libboost-serialization-dev. Since version 2.0.0 mlpack depends on the boost-serialization library
anshumitts has quit [Client Quit]
< aditya123>
zoq: Thanks, had missed libboost-serialization-dev. All set now
< zoq>
aditya123: Sounds great.
< acbull>
Hello, everyone. I've taken the machine learning course this semester, so I'm interested about your projects offered for GSoc, especially those about ML algorithm. Is it necessary to understand the algorithm and mathematics before handing in the proposal? Or I can list the study plan as the part of my proposal?
tanishq has joined #mlpack
Harsh has joined #mlpack
Harsh is now known as Guest15039
Guest15039 has quit [Quit: Page closed]
< zoq>
acbull: Hello, depending on the project, a willingness to dive into some literature on the topic and basic understanding of the c++ concepts and the related axioms are sufficient, so you can learn the rest over the summer. But there are some project you should have a full understanding of the description before applying, because you don't have the time to understand everything as time goes.
wasiq has joined #mlpack
tanishq has quit [Read error: Connection reset by peer]
tanishq has joined #mlpack
< tanishq>
#joing apertium
tanishq has left #mlpack []
acbull has quit [Ping timeout: 252 seconds]
< aditya123>
While doing make I get following error : undefined reference to `boost::unit_test::unit_test_main(bool (*)(), int, char**)'
< rcurtin>
aditya123: did you install libboost-test-dev?
< rcurtin>
that's not enough output for me to be able to help; please paste the full CMake configuration log, the CMakeCache.txt file, and the full build log
< rcurtin>
the CMakeCache.txt says "pastebin deleted"
< rcurtin>
also sorry I didn't ask for this before, but could you provide the output of 'VERBOSE=1 make'? that will show me what gcc is trying to link against
< rcurtin>
does the file /usr/lib/i386-linux-gnu/libboost_unit_test_framework.so exist?
< aditya123>
yes the file exists
< rcurtin>
this is not a problem I've seen before; I am thinking maybe something is odd about your setup because you configured and then installed some more libraries and then built...
< rcurtin>
could you try to just clean out the build directory and build again from scratch?
< rcurtin>
i.e., in the build directory, rm -rf *, then cmake ../, then make
< rcurtin>
and let's see if it does something different this time... :)
viswanadh has joined #mlpack
< acbull>
I 've noticed the reading list in the recommendation of project "Neuroevolution algorithms", but there indeed are a lot of materials. So I'm wondering where to start with
abaacaca has joined #mlpack
abaacaca has left #mlpack []
anveshi has joined #mlpack
< zoq>
acbull: Yeah, I absolutely agree with you that this two projects could be a great start to learn about various fundamental deep learning models from a practical perspective.
< zoq>
acbull: About the Neuroevolution algorithms project idea. A good start would be to read some paper about he neuroevolution algorithms mentioned in the project idea. You should also try to build mlpack and the code to communicate with the nes emulator.
< zoq>
acbull: The code to communicate with the emulator could be found here: https://github.com/zoq/nes. I'm about to write a readme for the nes emulation code, which should be finished by the end of the day.
< zoq>
acbull: I would start with the NEAT paper, but you don't have to read all or have to completely understand everything.
< zoq>
acbull: You can choose 2 or 3 algorithms you like and implement them over the summer. I guess NEAT is one of the coolest
< acbull>
When it comes to the emulation. What is the channel between our program and the channel? I mean, what is the output we can get from the emulation?
< rcurtin>
"NEAT is one of the coolest", funny play on words :)
< acbull>
Ok, that will be a good start~I will read the articles these days. Thanks for your precious recommendation, zoq~
< zoq>
acbull: matrix representation of the game state e.g. the postion of mario, the postion of enemies, the number of coins, number of lives blocks that you can jump on image (jpeg) of the current game screen
< zoq>
etc.
< zoq>
acbull: I make sure, to include some examples in the README
< zoq>
rcurtin: haha
< acbull>
So we have to let the program recognize the current state and make decisions for the next step, and all these work will be done by the program learning itself?
< tdgunes>
zoq: Hello, I am also interested in "Neuroevolution algorithms" project. I previously seen a lua implementation of a very primitive version of NEAT: https://www.youtube.com/watch?v=qv6UVOQ0F44. Do you think that someone without a background on neural network(not used but worked on several other ML algorithms) has a chance on that project?
< tdgunes>
zoq: It seems super interesting to work in the summer.
< zoq>
noooo I closed my browser ... hold on it takes a couple of minutes to reopen all tabs
< rcurtin>
zoq: I feel your pain, I've definitely done that too many times
< rcurtin>
I think there are usually extensions to help with this problem but I never think to install one, I always think "I'll just be more careful next time"
< rcurtin>
and then I do it again a few months later :)
nikola-miljkovic has quit [Remote host closed the connection]
< tdgunes>
zoq: do you want me to resend my previous messages ?
< zoq>
Okay, CPU back to 5%
< zoq>
tdgunes: Absolutely, we don't expect that anyone is already an expert in any of the topics, as long as you are willing to dive into the projects, it should be manageable. The NEAT algorithm is pretty straightforward to implement, so yeah, you can work on that project without a deep understanding about neural networks. Btw. the video is really neat :)
< zoq>
rcurtin: In fact I'll have OneTab installed, but maybe someone has to remind me to use that plugin.
< rcurtin>
:)
drktrx has joined #mlpack
mrbean has quit [Read error: Connection reset by peer]
< aditya123>
rcurtin: Working :D
< rcurtin>
aditya123: great, glad the solution was so straightforward :)
tanishq has joined #mlpack
acbull has quit [Ping timeout: 252 seconds]
vijaykrishnavans has joined #mlpack
vijaykrishnavans has quit [Client Quit]
minhazav_ has joined #mlpack
tanishq has quit [Quit: Leaving.]
varun has joined #mlpack
varun is now known as Guest33171
Guest33171 has quit [Client Quit]
kirizaki has joined #mlpack
< kirizaki>
Hello there everyone! So much people :)
sumedhghaisas has quit [Ping timeout: 252 seconds]
anveshi has quit [Quit: Page closed]
sumedhghaisas has joined #mlpack
sanket has quit [Ping timeout: 252 seconds]
ankur_ has joined #mlpack
ankur_ has quit [Client Quit]
ankur_ has joined #mlpack
< ankur_>
hi, I want to take work for Essential Deep Learning Modules for gsoc 2016
< ankur_>
how should I proceed
< ankur_>
?
viswanadh has joined #mlpack
sanket has joined #mlpack
viswanadh has quit [Remote host closed the connection]
ankur_ has quit [Quit: Page closed]
< zoq>
In the event ankur_ reads the irc logs
< zoq>
ankur_: Hello, glad you liked the idea. I think this project is a great opportunity to dive into some fundamental deep learning algorithms from a practical point of view. I've responded to a mail on the maillinglist that contains a bunch of papers and links you can read to get a better understanding about the essential network modules idea: https://mailman.cc.gatech.edu/pipermail/mlpack/2016-February/000738.ht
< zoq>
ml
< zoq>
okay, just 'finished' the NES readme, time to update the project idea :)
< tdgunes>
zoq: good to hear! I was waiting that to arrive :)
< zoq>
I'm thinking about modifying the CMake file so that OpenCv isn't necessary to build the code, but I'll guess I'll do that later.
tdgunes has quit []
tdgunes has joined #mlpack
sumedhghaisas has quit [Ping timeout: 252 seconds]
sanket has quit [Ping timeout: 246 seconds]
Manchv has quit [Ping timeout: 252 seconds]
kirizaki has quit [Quit: Konversation terminated!]
kirizaki has joined #mlpack
Guest42037 has quit [Ping timeout: 252 seconds]
tsathoggua has quit [Quit: Konversation terminated!]
kirizaki has quit [Quit: Konversation terminated!]