mlpack IRC logs, 2020-03-23
Logs for the day 2020-03-23 (starts at 0:00 UTC) are shown below.
--- Log opened Mon Mar 23 00:00:40 2020
07:49 < jenkins-mlpack2> Yippee, build fixed!
07:49 < jenkins-mlpack2> Project docker mlpack nightly build build #650: FIXED in 3 hr 35 min: http://ci.mlpack.org/job/docker%20mlpack%20nightly%20build/650/
07:58 -!- toluschr [~firstname.lastname@example.org] has joined #mlpack
08:00 -!- toluschr [~email@example.com] has quit [Client Quit]
08:25 < LakshyaOjhaGitte> hey do we have a link to the recurrent attention layer?
08:25 < LakshyaOjhaGitte> maybe its pr
08:39 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has joined #mlpack
09:03 -!- AbhiSaphire [firstname.lastname@example.org] has joined #mlpack
09:30 < PrinceGuptaGitte> Thanks a lot for your review @zoq . I've made all the necessary changes.
09:30 < PrinceGuptaGitte> I'll try to review it myself also more times so as to find more flaws.
09:57 -!- togo [~togo@2a02:6d40:34f8:8901:49e0:3aa4:c56f:b25] has joined #mlpack
10:57 -!- DestinyGamer [destinygam@gateway/shell/matrix.org/x-npvivxdstjhgarqx] has joined #mlpack
11:12 * DestinyGamer sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/JtvenYEDnxFmbKGwSUJWTwli >
11:20 -!- AbhiSaphire [email@example.com] has quit [Ping timeout: 240 seconds]
11:30 < LakshyaOjhaGitte> Hi @destinygamer No you are not too late. https://www.mlpack.org/gsoc.html might be helpful.
11:35 < LakshyaOjhaGitte> @rcurtin you online?
11:36 < LakshyaOjhaGitte> wanted some help with pr related to images
11:36 < LakshyaOjhaGitte> (edited) ... with pr related ... => ... with pr's related ...
11:37 < DestinyGamer> > Hi @destinygamer No you are not too late. https://www.mlpack.org/gsoc.html might be helpful.
11:37 < DestinyGamer> Thanks! May I know the handles of the mentors for the project?
11:40 < LakshyaOjhaGitte> I think you are looking for this https://github.com/mlpack/mlpack/wiki/SummerOfCodeIdeas.
11:41 < naruarjun[m]> @zoq is a mentor for RL
11:41 < JatoJoseph[m]> Hello guys, Pls, where can i submit my application?
11:41 -!- JatoJoe [firstname.lastname@example.org] has joined #mlpack
11:42 < JatoJoe> hello
11:42 < naruarjun[m]> @zoq is one for the RL project
11:42 < zoq> JatoJoe: Hello, via the GSoC page.
11:43 < naruarjun[m]> <JatoJoseph[m] "Hello guys, Pls, where can i sub"> You need to make a account at
11:44 < JatoJoseph[m]> <zoq "JatoJoe: Hello, via the GSoC pag"> okay...tahnks again zoq
11:45 < DestinyGamer> > @zoq is one for the RL project
11:45 < DestinyGamer> Thankss
11:50 < JatoJoe> zoq This https://github.com/mlpack/mlpack/wiki/Google-Summer-of-Code-Application-Guide does not provide a link fo rme to submit my application. Can you share teh submission link pls?
11:50 < Saksham[m]> You’ll need to go to official gsoc page and submit an application from there
11:50 < zoq> JatoJoe: https://summerofcode.withgoogle.com/ see the "STUDENTS APPLY NOW" button.
12:03 -!- AbhiSaphire [email@example.com] has joined #mlpack
12:05 < JatoJoseph[m]> Thanks Ma....I've seen it all...
12:05 < JatoJoseph[m]> <zoq "JatoJoe: https://summerofcode.wi"> Thanks Man...I done that'
12:05 -!- JatoJoe [firstname.lastname@example.org] has quit [Remote host closed the connection]
12:13 < M_slack_18> _<https://github.com/zoq/gym_tcp_api>_ I couldly understand what this exactly this is fornt exact
12:13 < M_slack_18> this exactly is for*
12:14 < M_slack_18> _<https://github.com/zoq/gym_tcp_api>_ I couldnt figure what this exactly does
12:14 < M_slack_18> using this
12:14 < M_slack_18> can we use gym environments in mlpack
12:27 < zoq> M_slack_18: Yes, that's bascially a wrapper around openai gym, so it can be used with mlpack (C++).
12:39 -!- favre49 [~email@example.com] has joined #mlpack
12:51 -!- AbhiSaphire [firstname.lastname@example.org] has quit [Remote host closed the connection]
12:53 -!- favre49 [~email@example.com] has quit [Quit: Lost terminal]
13:24 -!- M_slack_mlpack16 [slackml_39@gateway/shell/matrix.org/x-ijvrxedcuirtffvm] has joined #mlpack
13:29 -!- M_slack_mlpack16 is now known as LolitaNazarov[m]
13:30 < LolitaNazarov[m]> Hi! My name is Lolita Nazarov, I'm a 2nd year undergraduate at Stony Brook University majoring in Math + CS. I did an NSF-funded research project on Recurrent Neural Networks last summer, but my NSF-funded project for this summer got cancelled due to corona virus 😞 This is my first time ever contributing to an open source project! Do I stand a chance at being admitted to GSoC? I'm really interested in the
13:30 < LolitaNazarov[m]> Reinforcement Learning project - I have deep theoretical knowledge in that area. I hope I'm not too late either.
13:31 < LakshyaOjhaGitte> Hi @zoq
13:31 < LakshyaOjhaGitte> wanted to ask how does the recurrent attention layer work?
13:31 < Saksham[m]> Hi Lolita Nazarov
13:31 < Saksham[m]> <https://www.mlpack.org/community.html|https://www.mlpack.org/community.html> and <https://www.mlpack.org/gsoc.html|https://www.mlpack.org/gsoc.html> should you help you to get started
13:31 < LakshyaOjhaGitte> Hi, you are not too late.
13:32 < Saksham[m]> You can have a look <https://github.com/mlpack/mlpack/wiki/SummerOfCodeIdeas|https://github.com/mlpack/mlpack/wiki/SummerOfCodeIdeas> for the ideas
13:32 < LakshyaOjhaGitte> @zoq is there any link or pr link to the recurrent attention layer
13:34 < LolitaNazarov[m]> Thank you!!
13:37 < rcurtin> LolitaNazarov[m]: sorry to hear about the project being cancelled :(
13:37 < rcurtin> submissions are still open here, so please do feel free to put together an application. The ideas page has a link to an application guide too, which should be helpful
13:39 < LakshyaOjhaGitte> hey @rcurtin remember I asked you before regarding image handling in mlpack
13:39 < LakshyaOjhaGitte> I am confused as to where do we stand in regard to computer vision
13:43 < rcurtin> I don't remember, but I can try and answer a question. I'm not quite sure what your question is though
13:44 < rcurtin> you can do CV tasks in mlpack, you just have to "flatten" the images into Armadillo vectors (which is going to be what all the machine learning libraries do under the hood anyway to put them into neural networks or other ML algorithms)
13:44 < rcurtin> we have some image loading support via STB, and I know that people have connected the OpenCV matrix type with Armadillo and written converters (none are in mlpack though as we don't want to add an OpenCV dependency)
13:45 < rcurtin> I don't know, maybe that helps answer your question?
13:48 -!- Yashwants19 [firstname.lastname@example.org] has joined #mlpack
13:50 < Yashwants19> Hi rcurtin i m really interested for working in R binding as suggested by you and I mailed for the same.If you have some please take a look. Thank you
13:50 < Yashwants19> some time*
13:57 < Yashwants19> And for go binding release I have some questions regarding go get mlpack i have left some comments. Please share your opinions.Thank you
13:58 -!- Yashwants19 [email@example.com] has quit [Remote host closed the connection]
14:02 < rcurtin> Yashwants19: yeah, I saw that, just haven't had a chance to address it yet. Thanks for the responses, I hope to get to them shortly
14:04 < Param-29Gitter[m> Hey @rcurtin , I was planning to start working on support of reduction clause for different data-types Reference :https://software.intel.com/en-us/articles/openmp-40-new-features-supported-in-intel-compiler-160 but it requires support of OpenMP 4.0
14:05 < Param-29Gitter[m> and i guess mlpack currently works on OpenMP3.2
14:06 < Param-29Gitter[m> This would really be helpful in GSOC2020 profiling for parallelism and I wanted to know how to start working on the same...
14:06 < rcurtin> Param-29Gitter[m: reverse compatibility is important; I don't want to force new OpenMP features that aren't widely available
14:07 < rcurtin> visual studio doesn't even support OpenMP 3.0
14:08 < rcurtin> so, this significantly restricts what we can do in a platform-independent way
14:08 -!- Yashwants19 [firstname.lastname@example.org] has joined #mlpack
14:08 < Yashwants19> No problem rcurtin :)
14:09 -!- Yashwants19 [email@example.com] has quit [Remote host closed the connection]
14:10 < Param-29Gitter[m> Ok. So I guess I'll have to drop the idea completely as of now...
14:11 < Param-29Gitter[m> We can implement it maybe next year :)
14:12 < rcurtin> Param-29Gitter[m: or whenever MSVC gets their act together with OpenMP
14:13 < rcurtin> one option is to disable OpenMP support entirely on Windows because it's too old, but honestly I'm not very excited by that prospect
14:13 < rcurtin> (I'm not sure what Armadillo's approach is for that, might be worth looking at)
14:19 < LakshyaOjhaGitte> sorry I was gone for a bit, my main focus was what layers of new features you think are good for computer vision
14:26 < rcurtin> I don't have any thoughts on the matter, sorry; my focus is on other things
14:36 < LakshyaOjhaGitte> okay no problem, will hustle for a bit on it :)
14:36 < LakshyaOjhaGitte> (edited) ... no problem, will hustle for a bit on it :) => ... no problem
14:41 -!- toluschr [~firstname.lastname@example.org] has joined #mlpack
15:08 < Param-29Gitter[m> > `rcurtin on Freenode` Param-29 (Gitter): reverse compatibility is important; I don't want to force new OpenMP features that aren't widely available
15:08 < Param-29Gitter[m> I did not think about that. I did some reading turns out they do not show a lot of interest when it comes to supporting newer versions of OpenMP. They try supporting similar features but they have no plan in supporting this specific feature.
18:36 -!- M_slack_mlpack16 [slackml_40@gateway/shell/matrix.org/x-ngxtbjnmkmfcolxq] has joined #mlpack
18:45 -!- toluschr [~email@example.com] has quit [Quit: toluschr]
18:45 -!- M_slack_mlpack16 is now known as ZanHuang[m]
18:45 < ZanHuang[m]> Hi, this is Zan(<https://www.linkedin.com/in/pmixer/>), hope to know will the idea related to RNN based applications interesting for the community? I’m considering submitting the proposal for GSoC 2020 and having two applications prepared to propose: 1. RNN based recommender system for temporal networks: <https://github.com/srijankr/jodie> which is related to course project I’m currently working with and the instruct
18:45 < ZanHuang[m]> is the 1st author of jodie paper (kind of replacement for cf sample of mlpack). 2. Text-to-speech application: <https://github.com/NVIDIA/tacotron2> which is related to the work left when working for last employer. For both models, I plan to 0. reimplement in mlpack instead of relying on original pytorch version 1. update modules in mlpack if some feature needed has not been supported 2. try best to optimize the inference
18:45 < ZanHuang[m]> performance of the model and compare with other gpu/cpu implementations. Hope to know whether mentors would be interested and which one (jodie/tacotron) is preferred before writing the formal proposal. THX！
18:53 < rcurtin> ZanHuang[m]: hi there! I just looked at Jodie; are you at Georgia Tech?
18:54 < rcurtin> actually I see you are. small world! I am a GT grad and still live in Home Park :)
18:56 < rcurtin> anyway, the ideas seem cool, please do feel free to put together an application and submit
18:58 < naruarjun[m]> Hey, I had a doubt regarding PR #2311
18:58 < naruarjun[m]> The Q Learning test is failing for Linux Plain and Windows VS15 Plain
18:59 < naruarjun[m]> The test is compiling and running fine on my local build
19:00 < zoq> naruarjun[m]: Make sure to build with -DDEBUG=ON to enable bounds checks.
19:01 < naruarjun[m]> Yes I build with DEBUG=ON
19:03 < zoq> naruarjun[m]: Maybe it's related to, #2326, can you merge the current master brnach?
19:05 < ZanHuang[m]> THX Ryan, yes, gatech 😆 where u got Phd degree, I’ll elaborate on details and submit the proposal soon, BTW, is there one template for the proposal?
19:11 < sreenik[m]> freenode__slack_mlpack_U010J4P7MKJ[m]: Hey! I think this is what you are looking for https://github.com/mlpack/mlpack/wiki/Google-Summer-of-Code-Application-Guide
19:15 -!- AbhiSaphire [firstname.lastname@example.org] has joined #mlpack
19:15 < rcurtin> ZanHuang[m]: yeah, the application guide should be helpful, but there's no specific template
19:15 < naruarjun[m]> <zoq "naruarjun: Maybe it's related to"> Ok I’ll try that
19:16 < rcurtin> ZanHuang[m]: are you still at Tech now, or did they send all of the students home?
19:17 < ZanHuang[m]> campus closed, all using bluejeans for distance learning now. how’s it going with u, work from home?
19:20 < ZanHuang[m]> most students living nearby campus just stay in the room, guys living on the campus need to apply to stay but most should have left the campus and back to home.
19:28 < rcurtin> ZanHuang[m]: ah, ok, I don't get the emails from campus anymore :) I'm just working from home, going outside for a walk occasionally (but not today, it's rainy)
19:29 < rcurtin> I work for a remote company anyway, so working from home is nothing unusual for me :)
19:33 < ZanHuang[m]> take care, hope we all stay healthy guys~ guess WFH gives more time to develop mlpack lol
19:33 < rcurtin> yeah, true :)
19:53 -!- AbhiSaphire [email@example.com] has quit [Remote host closed the connection]
20:39 < M_slack_mlpack13> I had submitted a draft for reinforcement learning project. Looking forward to your comments and contributing to the organization.
20:44 < naruarjun[m]> I had also submitted a draft 2 days ago. Any comments are welcome.
20:54 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has quit [Quit: Leaving]
22:17 -!- nphamcs [firstname.lastname@example.org] has joined #mlpack
23:30 -!- togo [~togo@2a02:6d40:34f8:8901:49e0:3aa4:c56f:b25] has quit [Ping timeout: 272 seconds]
23:34 -!- togo [~togo@2a02:6d40:34f8:8901:49e0:3aa4:c56f:b25] has joined #mlpack
--- Log closed Tue Mar 24 00:00:42 2020