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/
< kris>
zoq: Do we have env.observation_space.low something equivalent gym module
< zoq>
kris: Not sure I get what you mean; do you like the equivalent of env.observation_space.low? I think it should be env.observation_space.boxLow
< zoq>
I wonder why I used boxLow instead of just low.
< zoq>
Btw. its defined in space.hpp.
< kris>
I saw that and was confused by the name boxLow i thought it was meant for the contrainer.
< kris>
but running std::cout << env.observation_space.boxLow.size() << std::endl; always give me 0
< zoq>
I commented line 109-112 in the worker.py file, because sometimes for some environments the worker just stuck at that point.
< zoq>
I didn't really looked into the issue because I thought, low and high isn't that much of interest.
< zoq>
If you include that lines it should work as expected.
< zoq>
I'll take a closer look at the issue later today.
< kris>
but i think that these values are important when you don't want to work in the continuous space but a discrete case like when you use q table solver. I will also look at the code and see if i can get it work
< zoq>
kris: ah, right. I think, I encountered some issues with the atari environments, not sure. But as I said it should work if you revert the change I did for line 109-112.
< zoq>
I guess you are using your own instance and not the public server?
govg has quit [Ping timeout: 260 seconds]
< kris>
zoq: Yes i am using my personal instance
< zoq>
kris: Okay, good, it's getting late here, time to get some sleep.
< kris>
zoq: Good Night. And thanks for all the help
kris has quit [Ping timeout: 260 seconds]
kris has joined #mlpack
vinayakvivek has joined #mlpack
govg has joined #mlpack
travis-ci has joined #mlpack
< travis-ci>
mlpack/mlpack#1874 (master - e0f6e97 : Ryan Curtin): The build passed.
mikeling|afk has quit [Quit: Connection closed for inactivity]
pyAddict has joined #mlpack
chvsp has quit [Ping timeout: 260 seconds]
govg has quit [Ping timeout: 260 seconds]
govg has joined #mlpack
govg has quit [Ping timeout: 268 seconds]
govg has joined #mlpack
kris has quit [Quit: Leaving.]
kris has joined #mlpack
kris has quit [Ping timeout: 260 seconds]
chvsp has joined #mlpack
vinayakvivek has quit [Quit: Connection closed for inactivity]
govg has quit [Quit: leaving]
pyAddict has quit [Quit: Page closed]
v_geta has joined #mlpack
< zoq>
hm, has to be another obscure failure with #830 on windows. I can't see anything in the log file, so this makes it even harder to solve the issue.
chvsp has quit [Ping timeout: 260 seconds]
kris has joined #mlpack
< zoq>
kris: Turns out that a connection timeout caused the problem, so I will revert the change I did.
hxidkd has joined #mlpack
hxidkd has quit [Client Quit]
Thyrix has joined #mlpack
< Thyrix>
Hello, is there any tutorial of ANN in mlpack?
< zoq>
Thyrix: Hello, not really, working on it, for now you best chance is to take a look at the tests:
< zoq>
rcurtin: Can you checkout the latest version of the mlpack.org repo?
hari_ has joined #mlpack
< kris>
zoq: Sorry i was not able to understand
hari_ has quit [Ping timeout: 260 seconds]
< kris>
Also i un commented the lines you said in worker.py file. And re- make the example.cpp files with no effect. It still shows 0 for the lowBox.size(). I am using a local server could that be a problem ??
< rcurtin>
seems like something is still hanging though, maybe there is some bug that only shows on sparc64
< rcurtin>
I guess that would make the whole effort of setting those up worthwhile then though!
< zoq>
I thought it's failing because it runs out of memory.
< rcurtin>
yeah, some of them are failing, I think that I need to reduce the number of executors on each system
< jarvis_>
zoq: I was thinking VGGNet should be in the mlpack arsenal if someone wished to utilise it. Because GoogLeNet is a drastic shift from conventional stacking of conv and pooling layers of the likes of LeNet/AlexNet architecture. So, it's not very easy to understand how to properly exploit it. There may be users who wish to utilise VGGNet instead. But you are right, GoogLeNet requires less parameters and less computation. I think I'll l
< jarvis_>
Inception v4 should be a good choice
< rcurtin>
jarvis_: looks like your first message got cut off at "I think I'll l", maybe there was more to the message?
< jarvis_>
I meant to say "I'll lean towards Inception v4"
< rcurtin>
:)
< jarvis_>
Hehe :D
jarvis_ has quit [Quit: Page closed]
v_geta has left #mlpack []
< zoq>
jarvis_: I get your point, I have to take a look at the actual numbers but isn't Inception V4 superior in terms of classication error and runtime? On the other side, the implementation is rather simple right? Convolution, Pooling, ... Linear, Softmax?
travis-ci has joined #mlpack
< travis-ci>
mlpack/mlpack#1880 (master - 40d4a39 : Ryan Curtin): The build passed.
< rcurtin>
hmm, ok, I remembered a different error on appveyor, so maybe I remembered incorrectly
< rcurtin>
I'm on a plane right now so I am not brave enough to try and load the appveyor logs :)
< zoq>
it's still building your change
< rcurtin>
yeah, I was waiting until I saw the result of that
< rcurtin>
and then I was going to decide whether I wanted to spend 5 minutes trying to load the page :)
< zoq>
:)
< rcurtin>
if it is the probabilities initialization like you pointed out, the issue is probably that the MSVC compiler isn't properly recognizing that the Armadillo operation can be cast to an arma::vec
< zoq>
std::vector<arma::vec>(1, arma::ones<arma::vec>(numObservations)/numObservations) not even sure that's correct
< zoq>
nice :)
< rcurtin>
awesome, we are part of GSoC! :)
< rcurtin>
soon the mailing list will explode :)
< rcurtin>
I think that that initialization is valid for std::vector, that should be std::vector(size_type n, value_type val)
< zoq>
:)
arunreddy has joined #mlpack
< zoq>
heading home now
< rcurtin>
ok, talk to you soon... maybe appveyor will be done by the time you get back
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
govg has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
florian__ has joined #mlpack
kesslerfrost has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
indra has joined #mlpack
indra_ has joined #mlpack
indra has quit [Client Quit]
indra_ is now known as indra
kesslerfrost has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
akasher has joined #mlpack
kesslerfrost has joined #mlpack
ravi_ has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
ravi_ has quit [Client Quit]
kesslerfrost has quit [Read error: Connection reset by peer]
keshav has joined #mlpack
kesslerfrost has joined #mlpack
mikeling|afk has quit [Quit: Connection closed for inactivity]
qwe_ has joined #mlpack
florian__ has quit [Ping timeout: 260 seconds]
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
qwe_ has quit [Ping timeout: 260 seconds]
akasher has quit [Ping timeout: 260 seconds]
walter__ has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
soumyadeep has joined #mlpack
sagar has joined #mlpack
vinayakvivek has joined #mlpack
walter__ has quit [Ping timeout: 260 seconds]
sagar is now known as Guest4693
< Guest4693>
guys i want to participate in gsoc what should i do to be in ....?
itsme__ has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
akasher has joined #mlpack
akasher has quit [Client Quit]
kesslerf_ has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
< Guest4693>
hi sir where should i start contributing for this organisation and if i face any problem so to whom i should contact
< rcurtin>
Guest4693: please start by consulting the mlpack website
< zoq>
soumyadeep: Okay, great, so I guess the next step is to explore the codebase, maybe you find an interesting issue on github, that you like to solve.
Guest4693 has quit [Ping timeout: 260 seconds]
kesslerf_ has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
kesslerfrost has quit [Read error: Connection reset by peer]
skk123 has joined #mlpack
kesslerfrost has joined #mlpack
soumyadeep has quit [Quit: Page closed]
kesslerfrost has quit [Read error: Connection reset by peer]
kesslerfrost has joined #mlpack
kesslerfrost has quit [Quit: kesslerfrost]
Sinjan_ has joined #mlpack
< Sinjan_>
I am really interested in working in this project. I have a strong foundation in c++. Besides I am involved in two machine learning projects over the past year. I will be really grateful if you can guide me where to start.
< zoq>
Sinjan_: Hello, and thank you for your interest, we have a special page that should provide a bunch of useful information: http://mlpack.org/gsoc.html
< Sinjan_>
I read the get introduced section. I am trying to install the stable version of mlpack in my Ubuntu. Can you help me out with the installation? The link is http://www.mlpack.org/files/mlpack-2.1.1.tar.gz
< rcurtin>
Sinjan_: note also you can just install mlpack via apt too
< Sinjan_>
I tried sudo apt-get install libmlpack-dev. But it seems its not available in my system's package manager.
< rcurtin>
maybe you are using an old version of Ubuntu?
< Sinjan_>
Ubuntu 14.04
< rcurtin>
yeah, unfortunately, it looks like mlpack has not been backported to that
< rcurtin>
if you upgrade to 16.04 or newer it is there, but that is probably too much work to be worth it :)
< rcurtin>
it's probably easier to just compile it at that point
< Sinjan_>
So, is there a way to install it on Ubuntu 14.04?
< rcurtin>
well, either install from source, try to find a ppa, or rebuild the .deb against 14.04... I think the first is the easiest
< rcurtin>
zoq gave you a link to the build directions
< Sinjan_>
I will try that.
Sinjan_ has quit [Ping timeout: 260 seconds]
Sinjan_ has joined #mlpack
arunreddy_ has joined #mlpack
arunreddy_ has left #mlpack []
Sinjan__ has joined #mlpack
Sinjan_ has quit [Ping timeout: 260 seconds]
< Sinjan__>
I tried installing it.
< Sinjan__>
But in the make step, I am getting a few errors.
< Sinjan__>
In file included from /home/user/mlpack-2.1.1/src/mlpack/../mlpack/core/tree/rectangle_tree/r_plus_tree_descent_heuristic.hpp:51:0, from /home/user/mlpack-2.1.1/src/mlpack/../mlpack/core/tree/rectangle_tree.hpp:36, from /home/user/mlpack-2.1.1/src/mlpack/../mlpack/methods/neighbor_search/neighbor_search.hpp:21, from /home/user/mlpack-2.1.1/src/mlpack/tests/aknn_test.cpp:12: /home/u
vinayakvivek has quit [Quit: Connection closed for inactivity]
< zoq>
Sinjan__: Can you use pastebin, gist.github.com or some other service to post your error?
< zoq>
Sinjan__: You can ignore the warnings, If I remember right there is an open PR that should solve all compiler warnings you see on your system.
< zoq>
Sinjan__: About the last error, my guess is your are not allowed to write into /usr/local/lib/ you can either use 'sudo make install' or specify another installation directory or just use the current build directory and skip the last step.
< Sinjan__>
Okay.
< Sinjan__>
sudo make install worked.
< zoq>
Sinjan__: Perfect
srishti_ has joined #mlpack
< Sinjan__>
I will now try out the command-line programs(man pages) on mlpack.
< zoq>
Sinjan__: Sounds good, if you need datasets you could take a look at: UCI Machine Learning Repository or mlpack already comes with some test dataset (src/mlpack/tests/data).
< Sinjan__>
Okay.
< Sinjan__>
In which directory are supposed to be to use these command line programs?
< zoq>
If you open a terminal and write e.g. 'mlpack_pca -h' it should work, if not /usr/local/bin/mlpack_pca -h should work for you.
srishti_ has quit [Quit: Page closed]
< Sinjan__>
The commands mlack_pca -h and /usr/local/bin/mlpack_pca -h are not working. Although, I noticed that all the command program files are there in /usr/local/bin/.
< rcurtin>
Sinjan__: I wonder if you have to set LD_LIBRARY_PATH
< rcurtin>
if it does, then you can just set LD_LIBRARY_PATH:
< rcurtin>
export LD_LIBRARY_PATH=/usr/local/lib
< rcurtin>
I believe there's a bit about this somewhere in README.md
< Sinjan__>
Okay
< Sinjan__>
It worked
< Sinjan__>
In the issues section, at first I want to work on an issue of easy difficulty level. But there is a separate tag P: major and P: minor. What are they about?