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/
< chopper_inbound4> Do we have a way to load .ckpt files in mlpack?
< bisakh[m]1> Hi zoq, I have submitted my draft proposal yesterday, please have a look, I would love to discuss/change/refine my idea accordingly.
< naruarjun[m]> Hey @zoq. I have also submitted a draft proposal on 22nd. Would appreciate if you could have a look and suggest changes that you think can help.
heisenbuug has joined #mlpack
heisenbuug has quit [Remote host closed the connection]
< jenkins-mlpack2> Project docker mlpack nightly build build #653: UNSTABLE in 3 hr 9 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/653/
heisenbuug has joined #mlpack
RudraPatil[m] has joined #mlpack
hemal[m] has joined #mlpack
TanayMehtaGitter has joined #mlpack
KimSangYeon-DGU has quit [Remote host closed the connection]
ImQ009 has joined #mlpack
heisenbuug has quit [Remote host closed the connection]
< zoq> chopper_inbound4: No, there is: https://github.com/sreenikSS/mlpack-Tensorflow-Translator that shoudl be helpful.
< zoq> bisakh[m]1: naruarjun[m]: Okay, I will take a loot once I have a chance, probably later today.
< sreenik[m]> kartickdutt18: Could you let me know about the status on grouped conv once?
< kartikdutt18Gitt> The gradient isn't working properly (tested locally), I can push the code if you want?
< sreenik[m]> No it's alright, take your time. I was just enquiring to make sure I didn't miss out in case it was completed
< kartikdutt18Gitt> I got the forward portion working, haven't tested backward yet but there is a test which has a full FFN that is trained on I think 4 from the mnist dataset. It took more 15 mins to fail. Once I am done I'll definitely tag you in the PR. Just incase, earlier it was implemented as separate layer in #2166
< sreenik[m]> Okay, sounds good
< sreenik[m]> bisakh: I took a quick glance at your proposal. It looks like you have combined two different project ideas in a single proposal. A better way to present it would be to submit two individual proposals if you are interested in working on either.
< sreenik[m]> I also feel that it would be great if the proposal's "title" matches the project idea name, especially because we have 2-3 ideas that are somewhat similar and changing a word or two makes it difficult to understand the idea for which it is submitted.
< chopper_inbound4> > chopper_inbound: No, there is: https://github.com/sreenikSS/mlpack-Tensorflow-Translator that shoudl be helpful.
< chopper_inbound4> Thanks @zoq I will take a look.
< bisakh[m]1> <sreenik[m] "I also feel that it would be gre"> Sure, I am changing it.
toluschr has joined #mlpack
< bisakh[m]1> <sreenik[m] "bisakh: I took a quick glance at"> Thank you so much @sreenik for having a look at my proposal. A few days back I was discussing with zoq about my idea of implementing those 2 modules, as currently we only have dcgan, wgan, wgan-GP implemented, there is a separate idea for pacGAN(which may get implemented), and I thought adding pix2pix can be something that we can boast of. I am confident about implementing
< bisakh[m]1> the 2 modules in this summer : ). Still, if you suggest I can put it in two proposals, but I think there will be a good amount of time left after completing any single module.
toluschr has quit [Quit: toluschr]
< Param-29Gitter[m> Hey @zoq, @rcurtin I did some low-level profiling of LinearSVM and PCA but results are not as I expected it to be. [This is a github repository where I updated my results](https://github.com/Param-29/mlpack-profiling). I tested these programs with 100k samples. These algorithms perform better when 'OMP_NUM_THREADS=1'. So I wanted to ask how do we as a organization decide which part of algorithm uses OpenMP and
< Param-29Gitter[m> which part doesn't.
< Param-29Gitter[m> Similar is the case with ANN training.
< Param-29Gitter[m> I am asking this question because I got confused. I had similar results in #2286 but we closed this PR thinking I couldn't achieve any speedup. But this is the case with many more algorithms.
< abhisaphire[m]> Hey rcurtin zoq . I have submitted a draft proposal please try to do a review when you get time, so that i would be able to make necessary changes within deadline...
< zoq> Ryan already mentioned this at some point, even if we don't get a chance to provde feedback in time, we can always reach out to you and ask for clarification.
< abhisaphire[m]> ok thanks zoq
< chopper_inbound4> Hi! zoq: What do you think about https://github.com/mlpack/mlpack/pull/2339 . Any suggestions from your side?
< metahost> Param-29Gitter[m: I think it's difficult to tell which algorithms might benefit from multiple threads (because fork/join overheads may vary) and result in a speedup. Profiling all of the algorithms would be a good indicator, you may write scripts to help automate the process to some degree; however, ymmv. :)
KimSangYeon-DGU has joined #mlpack
ImQ009 has quit [Read error: Connection reset by peer]
KarthikeyanR[m] is now known as KarthikeyanR[m]1
< KarthikeyanR[m]1> Hi zoq I have submitted my draft proposal please take a look. I would discuss and refine my ideas accordingly.
< KarthikeyanR[m]1> Also I am new to this community but I am a member contributing to the blender community. Could you tell me some issues so I could start contributing to the community?
< zoq> KarthikeyanR[m]1: Hello, there are some open issues listed on github, maybe there is something that you find interesting.
< pickle-rick[m]> The new GitHub mobile app looks pretty cool :)
< zoq> pickle-rick[m]: Haven't seen it yet.
travis-ci has joined #mlpack
< travis-ci> shrit/examples#19 (digit - ca9559d : Omar Shrit): The build is still failing.
travis-ci has left #mlpack []
KimSangYeon-DGU has quit [Remote host closed the connection]
KimSangYeon-DGU has joined #mlpack