ChanServ changed the topic of #mlpack to: "mlpack: a fast, flexible machine learning library :: We don't always respond instantly, but we will respond; please be patient :: Logs at http://www.mlpack.org/irc/
xiaohong has joined #mlpack
xiaohong has quit [Ping timeout: 276 seconds]
xiaohong has joined #mlpack
xiaohong has quit [Ping timeout: 245 seconds]
vivekp has quit [Ping timeout: 276 seconds]
xiaohong has joined #mlpack
xiaohong has quit [Ping timeout: 245 seconds]
jeffin143 has joined #mlpack
< jeffin143>
lozhnikov : Good Morning , yesterday night made some commit for 1904, can you just check and approve of the initial idea
< jeffin143>
Also , today I will clean up tf-idf , and also write blog by the eod. :)
< lozhnikov>
jeffin143: I think the initial idea is correct.
KimSangYeon-DGU has joined #mlpack
< sreenik[m]>
zoq: According to what we had discussed previously, I would be creating toString() functions for all the ann classes that I require right now (and maybe we can extend it to all of them later). In addition to that I also require functions to directly access the features of each layer (for eg: I need to know the kernel size, padding, etc of all the convolutional layers, I guess only inSize and outSize of all the layers can
< sreenik[m]>
be accessed right now). So would you suggest creating getter functions for all the layers I require right now (and maybe extend this later to all other layers too)? Also, another change is required and that is creating a getter function in the FFN class that would return the private object network. Please let me know if it is fine to go ahead with these and also if there is anything you would like to suggest regarding
< sreenik[m]>
these changes or would like to know about.
xiaohong has joined #mlpack
xiaohong has quit [Remote host closed the connection]
xiaohong has joined #mlpack
xiaohong has quit [Remote host closed the connection]
xiaohong has joined #mlpack
xiaohong has quit [Remote host closed the connection]
KimSangYeon-DGU has quit [Ping timeout: 260 seconds]
< zoq>
sreenik[m]: Implementing getter functions for various layer sounds reasonable.
< sakshamB>
ShikharJ: I am here
< Toshal>
ShikharJ: Yes me too.
< sreenik[m]>
zoq: Oh yes you had mentioned the visitor method the previous day too. Sounds good, I'll go ahead with it then
< rcurtin>
zoq: sreenik[m]: ToString(), blast from the past :)
< sreenik[m]>
Yeah :)
< rcurtin>
I think it has far more utility for neural network layers (where a user may want to print the network configuration) than it ever did for other types of mlpack models
< sreenik[m]>
Yes, since these models can get really big. Makes display work easier in projects and papers I suppos
< sakshamB>
ShikharJ: for now I have commented out a layer and so the build for minibatch discrimination should be passing. I have also made changes to padding layer.
KimSangYeon-DGU has joined #mlpack
ImQ009 has joined #mlpack
< akhandait>
rcurtin: Yup, it would be awesome to support printing of model configuration.
KimSangYeon-DGU has quit [Ping timeout: 260 seconds]
< zoq>
akhandait: I guess we could implement a summary function for the ffn/rnn class similair to keras.
KimSangYeon-DGU has joined #mlpack
vivekp has quit [Ping timeout: 244 seconds]
< ShikharJ>
sakshamB: Toshal: I should apologize for missing the meeting. I happened to move my apartment yesterday, and as a result, stayed awake late and woke up late :(
< ShikharJ>
sakshamB: That sounds good. So I think, you're currently working on Spectral Norm?
< ShikharJ>
Toshal: sakshamB: Let's spend the last two weeks of our time on getting the current PRs completed.
< ShikharJ>
sakshamB: In your case, since two of them (Padding and Minibatch) are close to be merge ready, and CGAN is there to be reviewed, and you're working on Spectral Norm, I think completing the current work till the last deadline should be the goal. We can work on StackGAN later on.
< ShikharJ>
Toshal: Same in your case, let's complete all work till LSGANs and maybe we'll look at other stuff post GSoC.
< ShikharJ>
Toshal: Also, please push out your work to blog posts, so that I can keep abreast of your work.
braceletboy has joined #mlpack
< braceletboy>
@zoq Have you made any updates to the mlplack-benchmarks directory?
< braceletboy>
I would like to add memory benchmarking scripts to it. But if there is work already done, I would be happy to build on it.
< braceletboy>
I don't think the the run.py script that you wrote, supports memory benchmarks
< zoq>
braceletboy: The PR should should be ready, I'll see if I can setup the jenkins job in next days.
< braceletboy>
Okay. Thank you.
< zoq>
braceletboy: About the memory benchmarks, ideally we could find a way to benchmark python code as well, in the previous version we used valgrind+massif to benchmark C++ code.
< zoq>
braceletboy: If you like to give this a shot, please feel free.
< braceletboy>
Okay. I will give it a shot. Thank you!
< zoq>
braceletboy: Sorry this took so long.
< braceletboy>
Ohh no. That's okay. I understand that you might be busy with GSOC mentorship
< braceletboy>
Have a nice day! :)
< zoq>
Thanks you too.
braceletboy has quit [Remote host closed the connection]
ImQ009 has quit [Quit: Leaving]
travis-ci has joined #mlpack
< travis-ci>
mlpack/mlpack#7813 (master - 1d17b8c : Ryan Curtin): The build has errored.
< abernauer>
rcurtin: Good news figured out problem the issue with getting the C++ code into an R session. Wrote the CLI utility code as well. Though upon using the functions I get runtime errors and then a core dump crashes the R session.
travis-ci has joined #mlpack
< travis-ci>
mlpack/ensmallen#427 (ensmallen-1.16.2 - fd51ff1 : Ryan Curtin): The build passed.