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/
chenzhe has joined #mlpack
trapz has quit [Quit: trapz]
< kris1>
zoq: I created the policy learning for the cart pole the program works fine but till 20 iterations is not learning very much
vinayakvivek has quit [Quit: Connection closed for inactivity]
eugenelee has joined #mlpack
< eugenelee>
hello
< eugenelee>
anybody here?
< eugenelee>
can you see me ?
eugenelee has quit [Client Quit]
vivekp has quit [Ping timeout: 260 seconds]
trapz has joined #mlpack
vivekp has joined #mlpack
trapz has quit [Quit: trapz]
trapz has joined #mlpack
niksoc has joined #mlpack
Trion has joined #mlpack
vivekp has quit [Ping timeout: 246 seconds]
vivekp has joined #mlpack
sagar has joined #mlpack
< sagar>
Hello
sagar is now known as Guest48786
Guest48786 has quit [Client Quit]
AndChat493044 has joined #mlpack
AndChat493044 has quit [Client Quit]
kartik has joined #mlpack
etotientz has quit [Ping timeout: 246 seconds]
vinayakvivek has joined #mlpack
sumedhghaisas has quit [Ping timeout: 240 seconds]
kartik has quit [Ping timeout: 246 seconds]
kartik has joined #mlpack
kartik has quit [Ping timeout: 246 seconds]
zoro_ has joined #mlpack
< zoro_>
hello ! i am interested in working on "Alternatives to neighborhood-based collaborative filtering". I am almost done with the reserach paper as mentioned so do we have to implement the algorithm using existing mlpack tools or something else ?
kartik has joined #mlpack
sicko has joined #mlpack
kartik has quit [Ping timeout: 246 seconds]
kartik has joined #mlpack
< rcurtin>
zoro_: the idea would be to implement it inside of mlpack; take a look at the mailing list archives to see other posts about the project that might help clarify things
kartik has quit [Ping timeout: 246 seconds]
cannon4 has joined #mlpack
travis-ci has joined #mlpack
< travis-ci>
mlpack/mlpack#2016 (master - 5b97d37 : Ryan Curtin): The build is still failing.
< zoq>
fysal: Have you seen: mlpack.org/gsoc.html and www.mlpack.org/involved.html; Also, definitely a good first step to take is to study the mlpack codebase closely and become familiar with its API.
< zoq>
:)
< rcurtin>
down to 22 flagged emails I need to address...
< zoq>
rcurtin: I can send you some more if you like :)
< Trion>
zoq: how can compiled executables carry the data they were trained on?
< rcurtin>
zoq: haha, it is a little overwhelming some days...
< rcurtin>
it'll probably take me most of the rest of the afternoon to get down to 0, if I am even able to put that much time into it today
< rcurtin>
headed out for lunch... back later
< zoq>
Trion: Maybe I misunderstood your comment, we still have to provide the model file (xml, csv, etc. ) that contains the parameters like weights, layer sizes etc.
< zoq>
Trion: But we don't reconstruct the network structure from the model file like the first layer is a linear layer followed by a ReLU layer etc, we could do that, but I'm not sure we should.
zoro_ has quit [Quit: Page closed]
cannon4 has quit [Ping timeout: 246 seconds]
< Trion>
But if we need to provide pre-trained models, then executables cannot be used as they get a fresh start every runtime
< Trion>
For pre-built models, executables are best option for sure
niksoc has quit [Quit: Page closed]
kartik has quit [Ping timeout: 246 seconds]
< zoq>
Trion: The executable should be able to load the model (pre-trained) e.g. take a look at the nbc executable that is able to save/load a model.
< zoq>
Trion: So let's say you have an executable for the GoogLeNet network you can train the network from scratch using something like: 'googlenet -i input.csv'
< zoq>
Trion: But you could also use the pre-trained model file to initialize the model and do what ever you like: 'googlenet -i input.csv -m pretrainedmodel.xml'
kartik has joined #mlpack
< Trion>
Oh I understand now :D I also saw that MLpack already has a serialization interface
< zoq>
Yeah, basically everything is already there, we just have to put it together.
ironstark has joined #mlpack
kartik has quit [Ping timeout: 246 seconds]
arunreddy has quit [Quit: WeeChat 1.4]
Trion has quit [Quit: CYA]
kartik has joined #mlpack
kartik has quit [Ping timeout: 246 seconds]
mikeling is now known as mikeling|afk
kartik has joined #mlpack
ironstark has quit [Ping timeout: 260 seconds]
kartik has quit [Ping timeout: 246 seconds]
sagar has joined #mlpack
sagar has quit [Client Quit]
sagarbhathwar has joined #mlpack
< sagarbhathwar>
Hello
< sagarbhathwar>
If I make add some functionality to a method, how can I verify if the method is doing as intended? Are there any tests available?
shikhar has joined #mlpack
< zoq>
sagarbhathw: Hello there, all tests are located in mlpack/src/mlpack/tests/; you can use mlpack_test to run all test or mlpack_test -t TestSuite to run a specific test suite you can even run a single test case with mlpack_test -t TestSuite/TestCase.