naywhayare changed the topic of #mlpack to: http://www.mlpack.org/ -- We don't respond instantly... but we will respond. Give it a few minutes. Or hours. -- Channel logs: http://www.mlpack.org/irc/
< jenkins-mlpack>
Ryan Curtin: Yet another reverse-compatibility fix for Armadillo.
Anand has joined #mlpack
< Anand>
Marcus : I added all the basic code required as suggested by you. Please have a look.
< Anand>
Btw, I will need to write some javascript code to use d3. Can we do that in the current reports code? You are generating some html there, right?
Anand has quit [Ping timeout: 246 seconds]
Anand has joined #mlpack
< Anand>
Marcus : How do you use the template file in reports code?
< Anand>
Also, is there any other way to access the metric string from db other than in the make_reports.py code? I ask this because we might want to keep the js code separate
< marcus_zoq>
Anand: We just fill the results into the gaps (templatr.py) that are marked with '%('.
< marcus_zoq>
Anand: I don't know exactly what you mean, you can everywhere access the databases.
< Anand>
Marcus : So basically you create a dict of the values to be inserted and then the '%' operator takes care of inserting the values in proper places, right?
< marcus_zoq>
Anand: Right, Btw. the code I sent you does not have any dependencies.
< Anand>
Yeah, I added the dependencies. I just used simplejson
< Anand>
No other module
< marcus_zoq>
Anand: Yeah, Right, I was talking about the integration into the make_reports.py file. You can use the code I sent you everywhere because there a no dependencies.
< Anand>
Ok. Right! I am integrating with the reports code, though.
< sumedhghaisas>
I am unable to explain it better :(
< naywhayare>
you'll have to call test<2>()
< sumedhghaisas>
I am doing that right now...
< sumedhghaisas>
any way to implement it like test(2)??
< naywhayare>
I don't think that's possible
< naywhayare>
not in C++03 at least
< naywhayare>
with C++11 you might be able to do something with constexpr to make it work
< naywhayare>
but I'm not sure exactly how you would do it
< sumedhghaisas>
constexpr??
< naywhayare>
it's a C++11 feature... it allows you to tell the compiler when a statement can be evaluated entirely at runtime
< naywhayare>
I'm not too familiar with it though
< sumedhghaisas>
there is contexpr meta-programming too... wow...
< sumedhghaisas>
naywhayare: constexpr meta-programming is not that effective...
< sumedhghaisas>
I wrote a sqrt program in 3 different ays
< sumedhghaisas>
*ways
< sumedhghaisas>
normal, constexpr meta-program and template meta-program...
< sumedhghaisas>
and computed their runtime...
< sumedhghaisas>
for number 1000000007
< sumedhghaisas>
normal - 352 microseconds
< sumedhghaisas>
constexpr - 332 microseconds
< sumedhghaisas>
template - 0 microseconds
< sumedhghaisas>
anyways complete incremental regularizations test is creating some problems... again regularization is not working...
< sumedhghaisas>
did you had a look at complete incremental learning??
< sumedhghaisas>
naywhayare: and can you please send me the changes done to CF module by siddharth?? I am unable update svn...
< sumedhghaisas>
git is working fine ... but not svn
< naywhayare>
sumedhghaisas: are you there? we should debug your subversion problems
< naywhayare>
Siddharth hasn't made any changes to the CF module
jbc__ has joined #mlpack
< jbc__>
I’m compiling with boost 1.55, and getting an error /usr/local/include/boost/smart_ptr/detail/shared_count.hpp:453:26: error: no member named 'get_untyped_deleter' in 'boost::detail::sp_counted_base'
< jbc__>
any ideas?
< naywhayare>
jbc__: can you give a backtrace to the mlpack code that causes the error?
< jbc__>
from CMAKE: [ 1%] Building CXX object src/mlpack/CMakeFiles/mlpack.dir/core/dists/discrete_distribution.cpp.o
< jbc__>
the configuration looked good
< jbc__>
full bt
< naywhayare>
okay; can you give the full error that gcc gave? it should be many lines long...
< sumedhghaisas>
yeah... I will do that tomorrow...
< sumedhghaisas>
campus placements are coming up... :(
< naywhayare>
how does a campus placement work? we don't have them here
< naywhayare>
what we have here are career fairs, where companies set up booths and students talk to them and give the companies resumes/CVs, and maybe get an interview if they are lucky
< sumedhghaisas>
seriously?? companies come to our campus for ... take group discussions ... interviews... coding rounds and etc...
< naywhayare>
I think maybe there are coding challenges sometimes, but usually it's just "wait in a line to talk to a recruiter"
< naywhayare>
no guarantee of a job or anything, though, the best one might do is to get an interview with a company
< sumedhghaisas>
thats bad...
< sumedhghaisas>
there is no guarantee here also... but at least you get the chance to sit for a coding round and show your skills...
< sumedhghaisas>
but sometimes there is aptitude round before that... in that non cs guys sit and makes our lives miserable...
< naywhayare>
so kind of like a programming competition? they give you little problems to solve and you submit a solution?
< sumedhghaisas>
umm... that happens in the second round... there is a online judge...
< naywhayare>
ah, okay
< sumedhghaisas>
in first round... they give sort of "what is the output" problems...
< sumedhghaisas>
puzzles...
< sumedhghaisas>
brainteasers...
< sumedhghaisas>
and all sorts of cs related questions... os... networking...
< sumedhghaisas>
pen and paper..
< sumedhghaisas>
if you get selected ... then second round of online coding...
< sumedhghaisas>
then 3 interview rounds...
< sumedhghaisas>
then a hr interview round...
< sumedhghaisas>
long process..
andrewmw94 has left #mlpack []
< naywhayare>
sounds stressful :(
< sumedhghaisas>
yeah... right now making my resume... very confusing...
< sumedhghaisas>
btw I wanted to tell you this long time back... I loved your site... ratml.org.. cool projects...
< naywhayare>
thanks :)
< naywhayare>
I haven't had time to finish most of my projects...
< sumedhghaisas>
especially lounge :)
< sumedhghaisas>
very nice ideas...
< sumedhghaisas>
some thing I could not understand...
< naywhayare>
I can explain whatever you didn't understand, if you are interested
< sumedhghaisas>
non cs stuff... so you have both electronics and computer background?? amazing...
< naywhayare>
ah, yeah. I was an electrical engineering undergraduate
< naywhayare>
my focus was signal processing... so... speech recognition and related fields
< naywhayare>
but... as I got deeper into signal processing I found out that what actually interested me was machine learning
< sumedhghaisas>
was it difficult shifting it to computers??
< naywhayare>
not really. I'd been playing with computers since I was 11 or 12
< naywhayare>
and I'd been doing development as a hobby for some years
< sumedhghaisas>
cause you know all the things they teach in computer science...
< naywhayare>
so really... I think maybe I chose wrong when I decided to do EE. should have been CS, I think :)
< sumedhghaisas>
great...
< sumedhghaisas>
I think I made a correct decision... I know nothing else...
< sumedhghaisas>
when did you learn C++?? that the most hated language in my campus...
< sumedhghaisas>
although its my favorite...
< naywhayare>
I learned C++ in high school for a project I was doing... that would have been 2004 I guess
< naywhayare>
but I didn't get very good at it until I started working on mlpack in 2009
< sumedhghaisas>
yes... I agree.. I started learning C++ when I made my first commit in september 2013 I guess...
< naywhayare>
:)
< sumedhghaisas>
I used to use JAVA before that...
< naywhayare>
I would still not say I am an expert
< naywhayare>
there are so many features I don't know about
< sumedhghaisas>
that the best thing about C++... everyday I learn something that I don' know... and C11.. they have introduced so many things in C++
< sumedhghaisas>
have you heard of MPL library of boost??
< sumedhghaisas>
its one of the coolest libraries I have ever seen...
< sumedhghaisas>
they have created vectors with variadic templates...
< naywhayare>
yeah, I think that one of the guys who worked on mlpack used it some time back
< naywhayare>
in src/mlpack/core/util/sfinae_utility.hpp, I think
< naywhayare>
or some file like that. he used it to detect if a class has a ToString() method
< sumedhghaisas>
ahh yes... I have read that file... we discussed it when we were discussing about converting it to std::string()