mlpack IRC logs, 2020-02-24

Logs for the day 2020-02-24 (starts at 0:00 UTC) are shown below.

February 2020
Sun
Mon
Tue
Wed
Thu
Fri
Sat
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
22
23
24
25
26
27
28
29
--- Log opened Mon Feb 24 00:00:00 2020
02:25 < PrinceGuptaGitte> I just saw now that convolution also take 2D inputs of dim (num samples, num features) and not 3d or 4d like(num samples, channels, width, height) like in keras.
02:25 < PrinceGuptaGitte> I was thinking it's a little odd, but it's also easier to maintain 2d tensors(mat)
02:26 < PrinceGuptaGitte> (Num features, num samples) *
04:39 < birm[m]1> Is masterblaster port 5005 down intentionally/known?
05:04 -!- Rudra [88e8d0be@136.232.208.190] has joined #mlpack
05:05 < Rudra> Hello,
05:08 -!- Rudra [88e8d0be@136.232.208.190] has quit [Remote host closed the connection]
05:08 -!- Rudra [88e8d0be@136.232.208.190] has joined #mlpack
05:08 -!- Rudra [88e8d0be@136.232.208.190] has quit [Remote host closed the connection]
06:13 < TanayMehtaGitter> I have a question, I know there are Neural nets in mlpack, but they should have something for automatic differentiation (like autograd for Pytorch) but I don't have any directions,
07:21 -!- aadarsh-as [75f287e2@117.242.135.226] has joined #mlpack
07:23 -!- aadarsh-as [75f287e2@117.242.135.226] has quit [Remote host closed the connection]
07:38 < GauravSinghGitte> Hi, @heytanay you can have look at [this](http://ensmallen.org/docs.html#differentiable-functions).
07:40 -!- gg95 [0e8bf0f7@14.139.240.247] has joined #mlpack
07:42 -!- gg95 [0e8bf0f7@14.139.240.247] has left #mlpack []
07:51 -!- Rudra [88e8d0be@136.232.208.190] has joined #mlpack
07:59 -!- Rudra [88e8d0be@136.232.208.190] has quit [Remote host closed the connection]
08:46 < TanayMehtaGitter> @gaurav-singh1998 thanks for the link! I had no idea something like this existed in cpp
11:00 * SaraanshTandonGi sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/JUnEAybQOcgnvjcMfwDNzvSW >
11:02 < SaraanshTandonGi> Also the exports to be done for LD_LIBRARY_PATH and PYTHONPATH in the end have been mentioned under the heading "Using mlpack without installing", whereas i needed to do it anyways.
11:42 -!- rohitsanj46 [01ba0c80@1.186.12.128] has joined #mlpack
12:32 * SaraanshTandonGi sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/eefjAPLvPERmTwAPekUlMaGJ >
12:37 < Nakul[m]> I think you don't need to master oop concept but basic are enough to get through the code and also.. template is very important.So I would suggest go through some online resources for that.
12:38 < Nakul[m]> Opting any easy issue and solving will clear lots of concept and workflow of the codebase
12:38 < Nakul[m]> And that's important
12:48 < SaraanshTandonGi> Thanks.
12:48 < SaraanshTandonGi> Could you please confirm the reproducability of the above problems cause i see someone has created an issue for a similar problem.
12:48 < SaraanshTandonGi> I would love to get started with a documentation PR to give me a morale boost.
12:50 -!- rohitsanj46 [01ba0c80@1.186.12.128] has quit [Remote host closed the connection]
12:58 < GauravSinghGitte> I don't know if this would be helpful but these are some of the articles I referred on the internet for the C++ concepts, [for C++ templates](https://www.codeproject.com/Articles/257589/An-Idiots-Guide-to-Cplusplus-Templates-Part-1), [for getting to know what is SFINAE](https://eli.thegreenplace.net/2014/sfinae-and-enable_if/), [for C++
12:58 < GauravSinghGitte> metaprogramming](https://discourse-production.oss-cn-shanghai.aliyuncs.com/original/3X/0/0/00329b35761ebb8cefec388d4423481dc836486f.pdf). I think you should start with the good first issues like [this](https://github.com/mlpack/mlpack/issues/2230) and learn side by side. Thanks.
13:00 < GauravSinghGitte> Hi, @saraansh1999 I don't know if this would be helpful but these are some of the articles I referred on the internet for the C++ concepts, [for C++ templates](https://www.codeproject.com/Articles/257589/An-Idiots-Guide-to-Cplusplus-Templates-Part-1), [for getting to know what is SFINAE](https://eli.thegreenplace.net/2014/sfinae-and-enable_if/), [for C++
13:00 < GauravSinghGitte> metaprogramming](https://discourse-production.oss-cn-shanghai.aliyuncs.com/original/3X/0/0/00329b35761ebb8cefec388d4423481dc836486f.pdf). I think you should start with the good first issues like [this](https://github.com/mlpack/mlpack/issues/2230) and learn side by side. Thanks.
13:24 < TanayMehtaGitter> I am facing 2 problems in building from source. 1. I have armadillo installed but it need version 8.400 and above. 2. When trying to install libensmallen-dev and libstb-dev, it says "unable to locate"
13:24 < TanayMehtaGitter> I'm using Ubuntu 18.04 and trying to build mlpack 3.2.2
13:26 < PrinceGuptaGitte> Hi @heytanay , uninstall libarmadillo-dev through apt-get as it installed lower version than required
13:27 < SaraanshTandonGi> remove the version of armadillo installed by apt-get and build it from source using their official documentation.
13:27 < SaraanshTandonGi> You can skip on libensmallen-dev and libstb-dev, CMake will automatically install them
13:27 < PrinceGuptaGitte> Just like he said then build the appropriate version from source
13:27 * SaraanshTandonGi sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/VIqriarbCrxxUrLzCKYcjFte >
13:28 < SaraanshTandonGi> I think it would be helpful to mention these small details in the documentation as they would avoid a lot of confusion for new people like me and would make the setup process a lot smoother
13:29 < PrinceGuptaGitte> Yeah I agree.
13:33 < SaraanshTandonGi> I'll try to write up a few modifications to the text. Things to include:
13:33 < SaraanshTandonGi> 1. Remove the armadillo library from the code snippet of apt-get, and add a line instructing to build it from source with a link.
13:33 < SaraanshTandonGi> 2. Remove libensmallen and libstb from the code snippet of apt-get informing about the flag that is by default set and handles the installation of these two automatically.
13:40 < TanayMehtaGitter> @saraansh1999 Good idea, Can I take on this small work? I want to get started contributing ASAP in any way possible. I will write both these points in `Building Mlpack from` source section on Github? What do you say?
13:40 < TanayMehtaGitter> (edited) ... Mlpack from` source section ... => ... Mlpack from source` section ...
13:40 < TanayMehtaGitter> @prince776 will try that rn
13:44 < SaraanshTandonGi> @heytanay actually it will be my first contribution too. So I would like to work on it. :)
13:44 < TanayMehtaGitter> @saraansh1999 Go ahead! All the best!
13:44 < SaraanshTandonGi> @heytanay Thanks a lot mate!
14:46 < TanayMehtaGitter> I'm doing `make` for mlpack on Linux and it's occasionally throwing some deprecation warnings (mostly with Numpy) is that normal?
14:48 -!- SakshamRastogiGi [gitterco_6@gateway/shell/matrix.org/x-qhrpryxrcvzcxrcb] has joined #mlpack
14:48 < SakshamRastogiGi> @heytanay I face the same warnings on Mac
14:56 < TanayMehtaGitter> @codeboy5 I believe it's okay as long as it's not throwing an error!
15:07 -!- susanna [5d2a1a0d@93-42-26-13.ip84.fastwebnet.it] has joined #mlpack
15:07 < susanna> Hello everyone!
15:10 < PrinceGuptaGitte> Hi everyone, can anyone who has good understanding of `ann` codebase help me with a doubt.
15:39 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has joined #mlpack
16:06 < PrinceGuptaGitte> Hi everyone, I'm facing an issue I can't resolve.
16:06 < PrinceGuptaGitte> The problem is I'm calling `boost::apply_visitor(outputWidthvisitor,network[0])` and I'm making sure there is getter `OutputWidth()` in the layer being visited. However it always returns 0. What could be the problem?
16:06 < PrinceGuptaGitte> I tried
16:07 < PrinceGuptaGitte> (edited) ... I tried => ... I tried to dodge it but using `ouputParameterVisitor` but the matrix it returns also shows 0 elements.
16:17 < KhizirSiddiquiGi> @prince776 : Many a times you need to train the layer one epoch at least to get the outputwidth parameter, if I am not wrong. I tried doing this summary thing quite a while ago.
16:18 < KhizirSiddiquiGi> Similar for `outputParameter` also.
16:18 < PrinceGuptaGitte> Oh thanks for the information. Do you have any idea what is the reason behind this
16:19 < PrinceGuptaGitte> Ohhh I understand now why it must be the case
16:20 < PrinceGuptaGitte> Thanks a lot. @khizirsiddiqui . I can finally progress.
16:20 < KhizirSiddiquiGi> :thumbsup: Do it bro, xD one really needs that feature sometimes.
16:21 -!- susanna [5d2a1a0d@93-42-26-13.ip84.fastwebnet.it] has quit [Remote host closed the connection]
17:16 -!- vroom_coder [~vroom_cod@106.51.84.215] has joined #mlpack
17:17 < vroom_coder> Hey, I am new here !
17:18 < KhizirSiddiquiGi> Hey vroom_coder!
17:18 < vroom_coder> I got to know about mlpack from GSOC and wish to contribute
17:19 < KhizirSiddiquiGi> You could look at these links for a start: https://www.mlpack.org/gsoc.html and https://www.mlpack.org/community.html
17:20 < KhizirSiddiquiGi> Ask here / mailing lists to clear if you have doubts. :)
17:24 -!- aadarsh-as [75dcafe4@117.220.175.228] has joined #mlpack
17:27 < vroom_coder> Thanks for the links! I will follow the steps.
17:27 -!- Pranshu54 [76b9ec74@118.185.236.116] has joined #mlpack
17:27 < vroom_coder> In the Essential Deep Learning Modules project, what are the other algorithms that are expected?
17:29 < aadarsh-as> hi i am a 2nd year undergrad student looking forward to work on reinforcement learning project under gsoc is there someone i can turn to for guidance
17:39 < vroom_coder> I want to contact Sreenik Seal. How can I do that?
17:42 < SriramSKGitter[m> @vroom_coder: Just ping him here with what you'd like to ask. He'll eventually get around to answering your query :)
17:45 -!- vroom_coder is now known as varun_agarwal
17:49 -!- varun_agarwal [~vroom_cod@106.51.84.215] has quit [Remote host closed the connection]
17:50 -!- varun_agarwal [~varun_aga@106.51.84.215] has joined #mlpack
17:52 -!- varun_agarwal [~varun_aga@106.51.84.215] has left #mlpack []
18:12 -!- aadarsh-as [75dcafe4@117.220.175.228] has quit [Ping timeout: 260 seconds]
18:14 -!- Pranshu54 [76b9ec74@118.185.236.116] has quit [Ping timeout: 260 seconds]
18:19 -!- varun_agarwal [~varun_aga@106.51.84.215] has joined #mlpack
18:20 -!- varun_agarwal [~varun_aga@106.51.84.215] has left #mlpack []
18:36 < kartikdutt18Gitt> Hi @aadarsh-as, take a look at https://www.mlpack.org/gsoc.html and https://www.mlpack.org/community.html .
18:39 -!- varun_agarwal [~varun_aga@106.51.84.215] has joined #mlpack
18:40 -!- varun_agarwal [~varun_aga@106.51.84.215] has left #mlpack []
18:51 -!- Pranshu54 [76b9ec74@118.185.236.116] has joined #mlpack
19:12 -!- Pranshu54 [76b9ec74@118.185.236.116] has quit [Remote host closed the connection]
19:18 -!- simplesit[m] [simplesitm@gateway/shell/matrix.org/x-mevmwbuyoenogmjs] has joined #mlpack
19:20 -!- ImQ009 [~ImQ009@unaffiliated/imq009] has quit [Quit: Leaving]
19:32 -!- Rudra [4f8e4c80@79.142.76.128] has joined #mlpack
20:02 -!- Rudra [4f8e4c80@79.142.76.128] has quit [Remote host closed the connection]
20:25 < zoq> vroom_code: as mentioned in the wiki, "it is probably not a great idea to contact mentors directly [...]", better to use the public channels
22:18 -!- travis-ci [~travis-ci@ec2-54-146-133-79.compute-1.amazonaws.com] has joined #mlpack
22:18 < travis-ci> coatless/ensmallen#8 (require-history-entry - 09291ce : James Balamuta): The build has errored.
22:18 < travis-ci> Change view : https://github.com/coatless/ensmallen/compare/2e9e1e7ddecf...09291cecf452
22:18 < travis-ci> Build details : https://travis-ci.com/coatless/ensmallen/builds/150385577
22:18 -!- travis-ci [~travis-ci@ec2-54-146-133-79.compute-1.amazonaws.com] has left #mlpack []
--- Log closed Tue Feb 25 00:00:01 2020