--- Log opened Thu Nov 15 00:00:03 2018 | ||
-!- HeikoS [~heiko@p4FED5ABB.dip0.t-ipconnect.de] has joined #shogun | 00:17 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 00:17 | |
-!- HeikoS [~heiko@p4FED5ABB.dip0.t-ipconnect.de] has quit [Ping timeout: 268 seconds] | 00:35 | |
-!- HeikoS [~heiko@p4FED5B4D.dip0.t-ipconnect.de] has joined #shogun | 10:48 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 10:48 | |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4412 merged by karlnapf | 10:55 |
---|---|---|
@sukey | [https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/b45243cf5afadcea5dee1df21284c2442ee01e98 by karlnapf | 10:55 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4405 merged by karlnapf | 10:57 |
@sukey | [https://github.com/shogun-toolbox/shogun] New commit https://github.com/shogun-toolbox/shogun/commit/6e42c52da66d995ca057ac7148682b1d4d77eea8 by karlnapf | 10:57 |
@sukey | [https://github.com/shogun-toolbox/shogun] Issue https://github.com/shogun-toolbox/shogun/issues/4413 closed by karlnapf | 11:07 |
-!- gf712 [c13cdcfd@gateway/web/freenode/ip.193.60.220.253] has joined #shogun | 11:45 | |
@HeikoS | gf712: yo! | 11:46 |
@HeikoS | gf712: so now that we have the attributes .... | 11:46 |
@HeikoS | we should think about how to best use this from shogun classes | 11:47 |
@HeikoS | I suggest we make a PR where we deploy this new masking idea to a few example classes | 11:47 |
gf712 | OK, would we still used the call from within the macro? | 11:47 |
@HeikoS | and then we document the idea and how to change it in an entrance task issue | 11:47 |
@HeikoS | yeah good question with the macro | 11:48 |
@HeikoS | I wonder whether we can overload it nicely | 11:48 |
@HeikoS | i.e. | 11:48 |
gf712 | OK, can give it a go with a simple example like Perceptron or LinearRegression | 11:48 |
@HeikoS | SG_ADD(pointer, name, mask)? | 11:48 |
@HeikoS | that should actually work or? | 11:48 |
gf712 | so is that the same as now but without the description? | 11:49 |
@HeikoS | ah sorry | 11:49 |
@HeikoS | description as well | 11:49 |
@HeikoS | SG_ADD(pointer, name, description, mask) | 11:49 |
@HeikoS | and then | 11:49 |
@HeikoS | SG_ADD(pointer, name, description) | 11:49 |
@HeikoS | (default mask) | 11:49 |
@HeikoS | those two should work | 11:49 |
@HeikoS | and then yes, a few examples | 11:50 |
gf712 | OK, and that replaces the current SG_ADD4? | 11:50 |
@HeikoS | yeah probably | 11:50 |
gf712 | and then a SG_ADD3 with the default? | 11:50 |
@HeikoS | the only thing is | 11:50 |
@HeikoS | that this will be hard to refactor | 11:50 |
@HeikoS | as currently an enum is passed | 11:50 |
gf712 | maybe.. Can give it a go and see how many errors come out.. | 11:50 |
@HeikoS | ok | 11:51 |
@HeikoS | I mean | 11:51 |
@HeikoS | it has to be done eventually | 11:51 |
gf712 | but the enums are binary so they are used usually in conditions no? So might just have to change the type in the declarations? | 11:51 |
@HeikoS | what do you mean? | 11:51 |
@HeikoS | i think most common are | 11:52 |
@HeikoS | SG_ADD(...) without any modelselection mention | 11:52 |
@HeikoS | and then SG_ADD(... MODEL_SELECTION_PARAMETER) | 11:52 |
gf712 | Oh right, that part, sorry | 11:52 |
@HeikoS | SG_ADD(&m_log_width, "log_width", "Kernel width in log domain", MS_AVAILABLE, GRADIENT_AVAILABLE); | 11:52 |
@HeikoS | GaussianKernel.cpp | 11:52 |
@HeikoS | uses both | 11:52 |
@HeikoS | SG_ADD(&properties, "properties", "Feature properties", MS_NOT_AVAILABLE); | 11:53 |
@HeikoS | Features.cpp | 11:53 |
@HeikoS | uses only one | 11:53 |
@HeikoS | I think SG_ADD3 is not defined right? | 11:53 |
@HeikoS | so omitting the properties completely wont compile | 11:53 |
gf712 | NO SG_ADD3 does not exist | 11:53 |
@HeikoS | ok | 11:53 |
@HeikoS | then you can let the compiler work for you | 11:53 |
@HeikoS | compiler errors | 11:53 |
@HeikoS | so SG_ADD5 is remove | 11:54 |
@HeikoS | SG_ADD4 uses the mask | 11:54 |
@HeikoS | and defaults to no property set | 11:54 |
@HeikoS | and SG_ADD3 doesnt exist | 11:54 |
@HeikoS | makes sense? | 11:54 |
gf712 | OK, just thinking how to set to no property | 11:55 |
@HeikoS | i think properties should always behave in a way that one explicitly has to add them | 11:55 |
@HeikoS | i.e. everything zero is the default | 11:55 |
@HeikoS | and that means | 11:55 |
gf712 | yes, I think that is what happens | 11:55 |
@HeikoS | parameter that is accessible via put/get | 11:55 |
gf712 | as in when calling AnyProperties enum class should be zero when instantiating | 11:55 |
@HeikoS | and thats all | 11:55 |
@HeikoS | yes, although I would be even more explicit | 11:56 |
@HeikoS | and AnyProperties doesnt have default ctor | 11:56 |
@HeikoS | ah sorry | 11:56 |
@HeikoS | it does | 11:56 |
@HeikoS | and then all zero | 11:56 |
@HeikoS | and that is used like you say | 11:56 |
@HeikoS | and then the author of the class can decide what is model parameter what has gradients etc | 11:56 |
@HeikoS | oh btw | 11:56 |
gf712 | sorry I mean enum class ParameterProperties | 11:56 |
@HeikoS | yes | 11:57 |
gf712 | what happens when it is instantiated? | 11:57 |
gf712 | is it all zero? | 11:57 |
@HeikoS | nothing is set | 11:57 |
@HeikoS | y | 11:57 |
gf712 | ok! | 11:57 |
@HeikoS | makes sense or? | 11:57 |
@HeikoS | gf712: oh btw | 11:57 |
@HeikoS | before you do that | 11:57 |
gf712 | yes, makes sense, I just wasn't sure if it was undefined behaviour | 11:57 |
@HeikoS | could you make a test for the getters | 11:57 |
gf712 | Yup, where would it make sense to put those? | 11:58 |
@HeikoS | so that we know the old API still works .... get_modelselection() | 11:58 |
@HeikoS | checking | 11:58 |
gf712 | btw should I add getters called is_*? | 11:58 |
@HeikoS | Any_unittest | 11:58 |
gf712 | if they return a boolean it makes more sense no? | 11:58 |
@HeikoS | gf712: yes | 11:58 |
@HeikoS | but why not general purpose | 11:58 |
@HeikoS | is(AnyParameterProperties) | 11:58 |
@HeikoS | is(AnyParameterProperties::MODEL) | 11:59 |
@HeikoS | sorry | 11:59 |
@HeikoS | bool has_property(enum class AnyParameterProperties) | 11:59 |
@HeikoS | and this one can be used for any property | 11:59 |
@HeikoS | like I call it as | 11:59 |
gf712 | true, that is more compact! | 12:00 |
@HeikoS | has_propertyAnyParameterProperties::MODEL) | 12:00 |
@HeikoS | or has_property(AnyParameterProperties::MODEL | AnyParameterProperties::GRADIENT) | 12:00 |
@HeikoS | or has_property(AnyParameterProperties::MODEL & AnyParameterProperties::GRADIENT) | 12:00 |
@HeikoS | all useful cases | 12:00 |
@HeikoS | ok cool | 12:00 |
@HeikoS | so then we have | 12:00 |
@HeikoS | 1) test the old API | 12:01 |
@HeikoS | 2) refactor SG_ADD | 12:01 |
@HeikoS | 3) solve all compile errors | 12:01 |
@HeikoS | i think 1) can be a separate PR | 12:01 |
@HeikoS | and then 2/3 can be one | 12:01 |
gf712 | cool! sound like fun :D | 12:01 |
@HeikoS | lol, well hopefully :D | 12:01 |
@HeikoS | and then next step would be to observe parameters is put | 12:01 |
gf712 | alright, Ill get to it! | 12:01 |
@HeikoS | cool! | 12:01 |
@HeikoS | Ill keep an eye on my emails | 12:02 |
gf712 | yup, can then discuss that with Viktor? | 12:02 |
@HeikoS | yeah the observe part we should discuss with him and Giovanni | 12:02 |
@HeikoS | as I think we should change the design a bit ... | 12:02 |
@HeikoS | but we will discuss | 12:02 |
@HeikoS | ah | 12:02 |
@HeikoS | another thing | 12:02 |
@HeikoS | ah nevermind | 12:02 |
@HeikoS | later :) | 12:02 |
gf712 | haha ok! | 12:03 |
gf712 | ill keep you updated! | 12:03 |
@HeikoS | cool man thx! | 12:03 |
@HeikoS | I am setting up the admin with the ATI btw | 12:03 |
@HeikoS | should be a contractor soon | 12:03 |
@HeikoS | then I can come visit London in December maybe | 12:03 |
@HeikoS | or latest in Jan | 12:03 |
gf712 | ah cool! so they pay for travelling? | 12:04 |
@HeikoS | oh and we wanna do a shogun workshop early next year, I think you should join as well! :) | 12:04 |
gf712 | yes, sounds good | 12:04 |
gf712 | where would it be? | 12:04 |
@HeikoS | gf712: see you! | 12:04 |
gf712 | see you! | 12:05 |
@HeikoS | gf712: maybe spain somewhere warm | 12:08 |
@HeikoS | and I might be in spain that time anyways | 12:08 |
@HeikoS | we usually have some funding for meetings | 12:08 |
@HeikoS | would be cool to get the ATI pay for your attending | 12:08 |
@HeikoS | I am sure we can ask :) | 12:08 |
gf712 | haha, yup that would be good! | 12:09 |
-!- travis-ci [~travis-ci@ec2-54-242-185-53.compute-1.amazonaws.com] has joined #shogun | 12:19 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/455413585 | 12:19 |
-!- travis-ci [~travis-ci@ec2-54-242-185-53.compute-1.amazonaws.com] has left #shogun [] | 12:19 | |
-!- travis-ci [~travis-ci@ec2-54-146-188-98.compute-1.amazonaws.com] has joined #shogun | 12:33 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/455413585 | 12:33 |
-!- travis-ci [~travis-ci@ec2-54-146-188-98.compute-1.amazonaws.com] has left #shogun [] | 12:33 | |
-!- travis-ci [~travis-ci@ec2-54-158-106-121.compute-1.amazonaws.com] has joined #shogun | 12:40 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/455414119 | 12:40 |
-!- travis-ci [~travis-ci@ec2-54-158-106-121.compute-1.amazonaws.com] has left #shogun [] | 12:40 | |
-!- travis-ci [~travis-ci@ec2-54-158-106-121.compute-1.amazonaws.com] has joined #shogun | 13:21 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun: https://travis-ci.org/shogun-toolbox/shogun/builds/455414119 | 13:21 |
-!- travis-ci [~travis-ci@ec2-54-158-106-121.compute-1.amazonaws.com] has left #shogun [] | 13:21 | |
-!- HeikoS [~heiko@p4FED5B4D.dip0.t-ipconnect.de] has quit [Ping timeout: 272 seconds] | 13:42 | |
-!- gf712 [c13cdcfd@gateway/web/freenode/ip.193.60.220.253] has quit [Ping timeout: 256 seconds] | 13:46 | |
-!- travis-ci [~travis-ci@ec2-54-163-94-123.compute-1.amazonaws.com] has joined #shogun | 15:08 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455494546 | 15:08 |
-!- travis-ci [~travis-ci@ec2-54-163-94-123.compute-1.amazonaws.com] has left #shogun [] | 15:08 | |
-!- travis-ci [~travis-ci@ec2-184-73-35-63.compute-1.amazonaws.com] has joined #shogun | 15:08 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455494546 | 15:08 |
-!- travis-ci [~travis-ci@ec2-184-73-35-63.compute-1.amazonaws.com] has left #shogun [] | 15:08 | |
-!- travis-ci [~travis-ci@ec2-54-163-94-123.compute-1.amazonaws.com] has joined #shogun | 15:24 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455494012 | 15:24 |
-!- travis-ci [~travis-ci@ec2-54-163-94-123.compute-1.amazonaws.com] has left #shogun [] | 15:24 | |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4414 opened by gf712 | 15:39 |
-!- travis-ci [~travis-ci@ec2-54-158-243-227.compute-1.amazonaws.com] has joined #shogun | 15:40 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455520780 | 15:40 |
-!- travis-ci [~travis-ci@ec2-54-158-243-227.compute-1.amazonaws.com] has left #shogun [] | 15:40 | |
-!- travis-ci [~travis-ci@ec2-54-147-192-193.compute-1.amazonaws.com] has joined #shogun | 16:07 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455505178 | 16:07 |
-!- travis-ci [~travis-ci@ec2-54-147-192-193.compute-1.amazonaws.com] has left #shogun [] | 16:07 | |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4415 opened by gf712 | 16:07 |
@sukey | [https://github.com/shogun-toolbox/shogun] Pull Request https://github.com/shogun-toolbox/shogun/pull/4415 closed by gf712 | 16:07 |
-!- gf712 [c13cdcfd@gateway/web/freenode/ip.193.60.220.253] has joined #shogun | 16:17 | |
-!- rcurtin [~ryan@knife.lugatgt.org] has quit [Ping timeout: 252 seconds] | 16:34 | |
-!- rcurtin [~ryan@knife.lugatgt.org] has joined #shogun | 16:40 | |
-!- travis-ci [~travis-ci@ec2-54-147-192-193.compute-1.amazonaws.com] has joined #shogun | 17:13 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455527526 | 17:13 |
-!- travis-ci [~travis-ci@ec2-54-147-192-193.compute-1.amazonaws.com] has left #shogun [] | 17:13 | |
-!- rcurtin [~ryan@knife.lugatgt.org] has quit [Ping timeout: 272 seconds] | 17:30 | |
-!- rcurtin [~ryan@knife.lugatgt.org] has joined #shogun | 17:36 | |
-!- gf712 [c13cdcfd@gateway/web/freenode/ip.193.60.220.253] has quit [Ping timeout: 256 seconds] | 17:40 | |
-!- travis-ci [~travis-ci@ec2-54-242-185-53.compute-1.amazonaws.com] has joined #shogun | 17:49 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455527526 | 17:49 |
-!- travis-ci [~travis-ci@ec2-54-242-185-53.compute-1.amazonaws.com] has left #shogun [] | 17:49 | |
-!- HeikoS [~heiko@p4FED5B4D.dip0.t-ipconnect.de] has joined #shogun | 19:05 | |
-!- mode/#shogun [+o HeikoS] by ChanServ | 19:05 | |
-!- travis-ci [~travis-ci@ec2-54-162-216-63.compute-1.amazonaws.com] has joined #shogun | 19:05 | |
travis-ci | it's Gil's turn to pay the next round of drinks for the massacre he caused in gf712/shogun: https://travis-ci.org/gf712/shogun/builds/455541908 | 19:05 |
-!- travis-ci [~travis-ci@ec2-54-162-216-63.compute-1.amazonaws.com] has left #shogun [] | 19:05 | |
@sukey | [https://github.com/shogun-toolbox/shogun] Issue https://github.com/shogun-toolbox/shogun/issues/4416 karlnapf added label: "good first issue" | 19:45 |
@sukey | [https://github.com/shogun-toolbox/shogun] Issue https://github.com/shogun-toolbox/shogun/issues/4416 karlnapf added label: "memory problems" | 19:45 |
@sukey | [https://github.com/shogun-toolbox/shogun] Issue https://github.com/shogun-toolbox/shogun/issues/4416 karlnapf added label: "bugfixing" | 19:45 |
@sukey | [https://github.com/shogun-toolbox/shogun] Issue https://github.com/shogun-toolbox/shogun/issues/4416 opened by karlnapf | 19:45 |
@sukey | [https://github.com/shogun-toolbox/shogun] Issue https://github.com/shogun-toolbox/shogun/issues/4416 karlnapf added label: "BUG" | 19:45 |
-!- HeikoS [~heiko@p4FED5B4D.dip0.t-ipconnect.de] has quit [Quit: Leaving.] | 19:56 | |
--- Log closed Fri Nov 16 00:00:04 2018 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!