verne.freenode.net 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/
gopalakr has joined #mlpack
< gopalakr>
hello #rcurtun, is there a binary for the sparse autoencoder or just the API interface ?
< rcurtin>
gopalakr: I didn't write the sparse autoencoder. if you don't see a sparse_autoencoder_main.cpp (which there isn't), then unfortunately you will have to write C++ to create a sparse autoencoder
< rcurtin>
it'd be hard to design a useful 'sparse_autoencoder' binary, so I think you'd be better off writing C++ anyway
< rcurtin>
(I have to get dinner... back later)
< gopalakr>
sure thanks
gopalakr has quit [Quit: Page closed]
< Rodya>
rcurtin do you guys have go bindings?
< Rodya>
i don't recall seeing any
< rcurtin>
Rodya: nope
< rcurtin>
I think Symantec will soon have me embarking on the automatic bindings project
< rcurtin>
since the internal teams want to use Python
< rcurtin>
so I'm hoping to write all the current binaries in some kind of "intermediate" C++ that can be parsed by a program that will output bindings for other languages
< Rodya>
iirc, go needs the extern C + cgo to work
< Rodya>
i wish C++ were easier to interact with from other languages
< rcurtin>
yeah, it is quite difficult... :(
< rcurtin>
it's not so horrible from Python
< rcurtin>
I wrote a quick python binding for some code I've been working on that I'll integrate soon (the code, not the binding); the main problem was getting the model to a way it could be represented by python
< rcurtin>
I ended up using boost::serialization to serialize it to an ostream, then the python class just contained the string representation... not the greatest solution..
< Rodya>
this pains me
< Rodya>
lol
< Rodya>
could look at using pickle instead of just ostream
< rcurtin>
yeah, I will at some point, but I was under deadline pressure :)
< rcurtin>
anyway, I'm off to bed (today I'm in Scotland, and it is now late)