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/
pie3 has joined #mlpack
hemal[m] has joined #mlpack
pie3 has quit [Ping timeout: 240 seconds]
ImQ009 has joined #mlpack
< sakshamb189[m]>
Hey himanshu_pathak are you there?
< himanshu_pathak[>
yr
< himanshu_pathak[>
Yes I am here
< himanshu_pathak[>
Hey sakshamb189: memory build was taking too much time I added some changes may be it should work for now
< sakshamb189[m]>
do you know why it was taking so long?
< himanshu_pathak[>
sakshamb189: Mostly when I try to handle `trainLabels` or `trainData` differently it start showing weird behaviour
< sakshamb189[m]>
I am not sure If I quite understand.
< himanshu_pathak[>
Even I am also not sure actually what the main cause is. Whenever I try to change something with these two variable it shows weird behaviour I tried to directly make copies without using for loop but still not working
< sakshamb189[m]>
I will take a look.
< himanshu_pathak[>
sakshamb189: Yup, we should wait for the current memory build to complete It should take 50 min atleast if worked normally
pie3 has joined #mlpack
< sakshamb189[m]>
I have left another review on the PR.
< sakshamb189[m]>
Would be great if you could commit those changes.
< sakshamb189[m]>
hmm also I think we should be serialising trainData and trainLabels. Let me know what you think
< himanshu_pathak[>
> I have left another review on the PR.
< himanshu_pathak[>
Thanks can I wait for current build to complete I will add them
< sakshamb189[m]>
yup sure we can wait.. but I think we do need to serialise the data
< himanshu_pathak[>
> hmm also I think we should be serialising trainData and trainLabels. Let me know what you think
< himanshu_pathak[>
Yeah I also think that
< himanshu_pathak[>
But trying check the behaviour of memory built without serializing'
< himanshu_pathak[>
* But trying check the behaviour of memory build without serializing'
< himanshu_pathak[>
Also I cannot use `size_t` instead of `int` beacuse to train smo we have to replace 0 with -1
< himanshu_pathak[>
External Api is same as other model user have to provide labels with 0 and 1 just I am using internal variable `trainLabel` to do mapping inside the implementation.
< sakshamb189[m]>
alright I think that should be fine
< himanshu_pathak[>
Anything else you want to discuss.
< sakshamb189[m]>
Let's just work on fixing the memory issue first. Did you find about using more than 2 classes with smo?
< himanshu_pathak[>
> Let's just work on fixing the memory issue first. Did you find about using more than 2 classes with smo?
< himanshu_pathak[>
Yeah I will try training multiple binary classifiers
< sakshamb189[m]>
alright. If there is nothing else you want to discuss right now then we can meet next week.
< himanshu_pathak[>
> alright. If there is nothing else you want to discuss right now then we can meet next week.
< himanshu_pathak[>
Yup I will try fixing memory issue today have a great week ahead :)
< sakshamb189[m]>
you too have a great week. Bye
< sakshamb189[m]>
himanshu_pathak: looks like the memory builds failed. Could you make the changes I suggested?
< himanshu_pathak[>
> himanshu_pathak: looks like the memory builds failed. Could you make the changes I suggested?
< himanshu_pathak[>
Yup doing that
< sakshamb189[m]>
also serialize the parameters
pie3 has quit [Ping timeout: 264 seconds]
< himanshu_pathak[>
sakshamb189: Done
pie3 has joined #mlpack
< sakshamb189[m]>
I have left a comment.
< sakshamb189[m]>
but let's wait for the build to finish first.
< himanshu_pathak[>
Yup, wait for the builds to complete :)
pie3 has quit [Max SendQ exceeded]
pie3 has joined #mlpack
pie3 has quit [Read error: Connection reset by peer]
pie3 has joined #mlpack
< walragatver[m]>
birm: jeffin143: I would not be a able to attend today's meet.
< walragatver[m]>
So I am just writing the things I want to highlight.
< walragatver[m]>
I think #7 is quite ready now. Except one or two comments that are remaining. I will give a final look tomorrow and will merge it in.
< jeffin143[m]>
Ok
< walragatver[m]>
jeffin143: Have you given any thoughts on the ensmallen callback design?
< jeffin143[m]>
Yes I have given
< jeffin143[m]>
I will write some code
< jeffin143[m]>
And show it to you off possible
< walragatver[m]>
jeffom143: Okay. Anything you would like to discuss?
< walragatver[m]>
jeffin143: .
< jeffin143[m]>
No all good from my side
< jeffin143[m]>
I will debug the folder issue
< jeffin143[m]>
@zoq : if you are there could I ssh into the machine you have Himanshu
< jeffin143[m]>
Just for debugging purpose if possible ???
< jeffin143[m]>
Gave* Himanshu
< walragatver[m]>
jeffin143: Yeah I also sent a link on the markdown issue just check it out
< walragatver[m]>
If you need ssh I have one login
< walragatver[m]>
Just a sec
< walragatver[m]>
I will sent it to you by mail
< jeffin143[m]>
Ohk thanks that would help.me a lot
< jeffin143[m]>
Thanks :)
< jeffin143[m]>
I will pakka take the time out and clear at least the folder issue once and for all
< jeffin143[m]>
So that we can get that out of our way
< walragatver[m]>
Okay it's not that important at this moment. So it's up to you
< sakshamb189[m]>
himanshu_pathak: how long is the memory check supposed to take?
< jeffin143[m]>
Yeah , also need to debug small embedding re error
< walragatver[m]>
jeffin143: Have you seen the solution I mentioned in comment in #7 about resolving _preprocess directory problem?
< jeffin143[m]>
Yes I saw
< walragatver[m]>
jeffin143: No sorry it was about _templogs sorry
< jeffin143[m]>
Yes I was about to say
< walragatver[m]>
I think _preprocess directory also got resolved yesterday just let me check
< walragatver[m]>
Oh yeah _preprocess also getting removed now.
< walragatver[m]>
jeffin143: Okay so it's great
< walragatver[m]>
Just fix the _templogs directory problem. And we are good to go
< jeffin143[m]>
I think we should just add some time lag
< jeffin143[m]>
Swap files or temp files are created when something is being written to it
< jeffin143[m]>
I will try out
< jeffin143[m]>
Just before logging off
< jeffin143[m]>
It would be great if you could create
< jeffin143[m]>
A user and share ssh
< jeffin143[m]>
Access
< walragatver[m]>
I cannot create a user
< walragatver[m]>
But I could share my credentials
< walragatver[m]>
jeffin143: Tagging you
< jeffin143[m]>
Ok sure that would work
< walragatver[m]>
> I think we should just add some time lag
< walragatver[m]>
Swap files or temp files are created when something is being written to it
< walragatver[m]>
Okay or just try out the work around I mentioned in #7's comment.
< jeffin143[m]>
Ok
< jeffin143[m]>
I will do
< walragatver[m]>
jeffin143: My Savannah login is not working.
< jeffin143[m]>
Umm ok
< walragatver[m]>
rcurtin: I am not able to ssh to Savannah is it down?
< walragatver[m]>
jeffin143: birm: Signing off take care.
< jeffin143[m]>
Sure
< jeffin143[m]>
Bye
< himanshu_pathak[>
sakshamb189: It take around 50 min generally but took 2hrs 16min strange
pie3 has quit [Ping timeout: 265 seconds]
< rcurtin>
walragatver[m]: we lost connection to savannah some time back :(
< rcurtin>
that was one of the systems that was hosted by Symantec, and I no longer have any touch with anyone in the company
< rcurtin>
it seems like they turned them off
< rcurtin>
sorry :(
< abernauer[m]>
rcurtin: My sunday plans got canceled so I have more time for code review :)