Open in new window / Try shogun cloud
--- Log opened Tue Aug 23 00:00:17 2016
-!- OXPHOS [4441a916@gateway/web/freenode/ip.] has joined #shogun00:28
@wikingOXPHOS, ping02:40
OXPHOSwiking: hey02:41
@wikingthese are explicitly moved to public02:43
@wikingin develop02:43
OXPHOSwiking: ohhhh i see. sure please move them out. I think the friend class can also be removed02:44
@wikingand the other patch i needed to fix02:45
@wikingthat in euclideandistance the precompute becomes like this
@wikingi.e. no linalg::dot involved02:47
@wikinganyhow thye are really minor changes02:47
@wikingso if you dont mind02:47
@wikingi'll do now a forcepush to the branch02:47
@wikingthat way the branch gets rebased on develop02:48
OXPHOSwiking: sure please go ahead!02:51
@wikingok pushing!02:51
-!- shogun-notifier- [] has joined #shogun02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * f4cc9b6 / / (16 files):
shogun-notifier-shogun: working version02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 7935f14 / / (27 files):
shogun-notifier-shogun: set default cpubackend02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * d291a7e / src/shogun/mathematics/linalgrefactor/ (5 files):
shogun-notifier-shogun: const02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * dad5a09 / src/shogun/mathematics/linalgrefactor/ (10 files):
shogun-notifier-shogun: float3202:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 19f155b / benchmarks/linalg_refactor_benchmark.cpp:
shogun-notifier-shogun: CPU and GPU dot02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * e60796d / / (20 files):
shogun-notifier-shogun: fix style02:56
shogun-notifier-shogun: Linalg-SGVector-add02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 1b28003 / / (4 files):
shogun-notifier-shogun: Remove linalg::dot02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 8573dbb / src/shogun/mathematics/linalg/LinalgBackendViennacl.h:
shogun-notifier-shogun: Linalg_ViennaCL_typecast_helper_method02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * bda14ca / tests/unit/mathematics/linalg/operations/
shogun-notifier-shogun: fix viennacl unittest02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 3011d59 / src/shogun/mathematics/linalg/ (3 files):
shogun-notifier-shogun: add DEFINE_FOR_ALL_PTYPE macro02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 3071eb1 / src/shogun/ (6 files):
shogun-notifier-shogun: update linalg for SGMatrix compatibility02:56
shogun-notifier-shogun: - change gpu_vector to gpu_ptr02:56
shogun-notifier-shogun: - change SGVector<> to Container <> in macros02:56
shogun-notifier-shogun: - left methods implementations intact as some methods are not SGMatrix02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 4afe07f / src/shogun/mathematics/linalg/LinalgBackendGPUBase.h:
shogun-notifier-shogun: update LinalgBackendGPUBase macros02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 0c66fdf / / (8 files):
shogun-notifier-shogun: linalg-sgvector-add-inplace02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 52a1e9f / / (6 files):
shogun-notifier-shogun: linalg-sgmatrix-add02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 3c6c849 / / (6 files):
shogun-notifier-shogun: linalg-sgvector&sgmatrix-scale02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * b61e283 / / (7 files):
shogun-notifier-shogun: linalg-sgvector&sgmatrix-max02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * cb815ea / / (6 files):
shogun-notifier-shogun: linalg-sgvector and sgmatrix-set const02:56
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 0b25451 / / (4 files):
shogun-notifier-shogun: linalg-sgvector-range_fill02:56
OXPHOSwiking: thanks!02:59
-!- travis-ci [] has joined #shogun03:18
travis-ciit's OXPHOS's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun:
-!- travis-ci [] has left #shogun []03:18
@wikingOXPHOS, ok so just a quick follow up on the issue03:43
@wikingwith the rotational matrix03:43
@wikingOXPHOS, we could change a bit the function declaration of that api03:44
@wikingvoid rotational_matrix(int32_t size, SGMatrix<T>& matrix);03:44
@wikingthat way we wouldnt have any problem to infer things03:45
-!- OXPHOS [4441a916@gateway/web/freenode/ip.] has quit [Ping timeout: 264 seconds]04:25
-!- sanuj [~sanuj@] has joined #shogun05:22
-!- OXPHOS [4441a916@gateway/web/freenode/ip.] has joined #shogun05:48
OXPHOSwiking: yes with SGMatrix<T> in param it can infer05:49
OXPHOSsanuj hey sry cannot do pm for some reason..05:51
sanujOXPHOS, did you submit your work on gsoc website?05:52
OXPHOSsanuj: not yet..close05:52
sanuji'm submitting it like this05:53
sanujdo you think it's fine?05:53
-!- shogun-notifier- [] has quit [Quit: transmission timeout]05:56
OXPHOSsanuj: I don't know the rules - but I think as long as you have enough PRs it should be fine?05:56
OXPHOSsanuj: I checked some report people shared on facebook and they're kinda elabroated.05:57
OXPHOSsanuj: I have one last part left..but it's more or less like this:
sanujoh wow05:58
sanuji should improve mine :D05:58
@wikingi wouldn't sweat it so much05:59
@wikingalthough i dont know how it is now05:59
@wikingback in my time this wasn't so strict of the code submission05:59
sanujwiking, ohkay05:59
@wikingi mean i would say both are fine06:00
@wikingsanuj: yours is very clean and simple06:00
@wikingOXPHOS: did an essay :)06:00
OXPHOSwiking: :D like README06:00
sanujoh, i guess i'll just submit06:01
sanuji can edit the gist later too06:01
sanujif required06:01
OXPHOSu r right06:02
sanujOXPHOS, are you back in the US?06:02
@wikingoh man merging all these feature branches is going to be fun :)06:03
OXPHOSsanuj: yes..but in california now. will go back to east coast by the end of this week06:03
sanujOXPHOS, remember i got an intern at JHU06:03
OXPHOSsanuj: yes!06:03
sanujthose people are taking a lot of time with the paperwork required for me to apply for visa06:03
sanujit has been almost a month06:03
OXPHOSwiking: I need to revise some stuff based on the tag/plugin change06:04
@wikingthat's alright06:04
OXPHOSsanuj: you're running out of time?06:04
sanujOXPHOS, they told me the delay is due to some US Federal Employment laws that are set to go into place on December 1, 201606:05
sanuji suspect the law is this:
sanujOXPHOS, you think they might have second thoughts and cancel the internship invitation as they won't be able to fund me with the new law in place?06:06
sanujbecause in that case i should start looking for new research internships06:06
OXPHOSsanuj: sorry I had no clue about it..they shouldn't, but who knows..I have a friend who got kicked out of graduate school after 1st year because his P.I. "suddenly run out of money" after taking him06:08
OXPHOSsanuj: so I'll suggest you start looking for new it's gonna take a while anyway06:08
OXPHOSdo they mean they'll start signing the paperwork after Dec 1st...?06:09
sanuji don't think so06:14
sanujOXPHOS, the mail said: they need confirmation from the prof about my appointment, after the knowledge of this new law.06:15
OXPHOSsanuj: oh so I guess you can harass the prof for the confirmation letter and see what he thinks?06:17
sanujOXPHOS, seems like the prof is very busy, people working in the lab did my interview and everything else06:18
sanuji don't like applying for research interns, most of the profs don't even read my mails :)06:19
OXPHOSsanuj: i got no reply from industry either...I guess it's because of my major but I was applying for bioinfo06:20
OXPHOSsanuj: but you have to e-mail the prof since he decides to take you06:20
OXPHOSsanuj: my experience is they'll blame you if you missed something because you're too "polite" or patient with them XD06:21
sanujokay i'll send a mail today06:22
OXPHOSgood luck!06:22
sanujOXPHOS, thanks! :D06:22
-!- sanuj [~sanuj@] has quit [Ping timeout: 264 seconds]07:10
-!- sanuj [~sanuj@] has joined #shogun07:56
Saurabh7uhm OXPHOS your link looks broken08:17
OXPHOSSaurabh7: thanks! I changed the title XD
Saurabh7ah isee careful with that :)08:34
-!- OXPHOS [4441a916@gateway/web/freenode/ip.] has quit [Quit: Page closed]09:21
-!- sanuj [~sanuj@] has quit [Ping timeout: 244 seconds]09:26
-!- sanuj [~sanuj@] has joined #shogun10:03
-!- shogun-notifier- [] has joined #shogun10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 196eb1e / / (3 files):
shogun-notifier-shogun: Remove linalg::dot10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 3d5816e / src/shogun/mathematics/linalg/LinalgBackendViennacl.h:
shogun-notifier-shogun: Linalg_ViennaCL_typecast_helper_method10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 70fdbe7 / tests/unit/mathematics/linalg/operations/
shogun-notifier-shogun: fix viennacl unittest10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * aac7cdf / src/shogun/mathematics/linalg/ (3 files):
shogun-notifier-shogun: add DEFINE_FOR_ALL_PTYPE macro10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 3925825 / src/shogun/ (6 files):
shogun-notifier-shogun: update linalg for SGMatrix compatibility10:11
shogun-notifier-shogun: - change gpu_vector to gpu_ptr10:11
shogun-notifier-shogun: linalg-sum_no_diag10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 56db07a / src/shogun/mathematics/linalg/ (3 files):
shogun-notifier-shogun: add return value/fix msgs/remove white lines10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 2423b14 / / (5 files):
shogun-notifier-shogun: linalg-matrix block sum10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 138e3fe / / (6 files):
shogun-notifier-shogun: linalg-sgvector-add-inplace10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 3892385 / / (6 files):
shogun-notifier-shogun: linalg-sgmatrix-add10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 65af379 / / (6 files):
shogun-notifier-shogun: linalg-sgvector&sgmatrix-scale10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 3e2c334 / / (7 files):
shogun-notifier-shogun: linalg-sgvector&sgmatrix-max10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 8e481f8 / / (6 files):
shogun-notifier-shogun: linalg-sgvector and sgmatrix-set const10:11
shogun-notifier-shogun: OXPHOS :feature/linalg_refactor * 86708a8 / / (4 files):
shogun-notifier-shogun: linalg-sgvector-range_fill10:11
-!- sanuj [~sanuj@] has quit [Ping timeout: 252 seconds]10:14
-!- sanuj [~sanuj@] has joined #shogun10:23
-!- travis-ci [] has joined #shogun10:59
travis-ciit's OXPHOS's turn to pay the next round of drinks for the massacre he caused in shogun-toolbox/shogun:
-!- travis-ci [] has left #shogun []10:59
-!- sanuj [~sanuj@] has quit [Ping timeout: 260 seconds]11:01
-!- sanuj [~sanuj@] has joined #shogun11:35
-!- shogun-notifier- [] has quit [Quit: transmission timeout]13:11
-!- sanuj [~sanuj@] has quit [Ping timeout: 276 seconds]14:03
-!- sanuj [~sanuj@] has joined #shogun14:58
-!- sanuj [~sanuj@] has quit [Ping timeout: 252 seconds]15:19
-!- Benni_ [86020a1f@gateway/web/freenode/ip.] has joined #shogun15:31
Benni_Hi can anybody help me with GP-Classification+MKL? More specifically, I'm not sure if the  kernelweights have been learned.15:33
Saurabh7Benni_: use get_subkernel_weights() maybe15:55
Benni_Im using the VB-inference + gradient based model selection framework (similar to the last chapter in the VB-notebook). However, the subkernel weights are all 1.0 before and after model selection (Im using 3 subkernels). Im not sure whether I have to explicitly specify the kernel weights as hyperparameters, or if they are autodetected and optimized during model selection. If I have to explicitly specify the weights, how do I do this (in 16:24
-!- sanuj [~sanuj@] has joined #shogun17:33
-!- Benni_ [86020a1f@gateway/web/freenode/ip.] has quit [Quit: Page closed]17:38
-!- Hieronymous [~sudarshan@] has joined #shogun17:44
-!- Hieronymous [~sudarshan@] has quit [Quit: Leaving]18:53
-!- sanuj [~sanuj@] has quit [Remote host closed the connection]19:49
--- Log closed Wed Aug 24 00:00:19 2016