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/
< walragatver[m]>
jeffin143 : Let's implement removing the directory and it's files mechanism in CMake. It would avoid Unix system calls so the fix will also work for windows in future. Also can we provide a CMake option which if ON it will remove eventfiles and directories generated during testing or else when off it will not remove it?
< walragatver[m]>
jeffin143: Also let's change the directory structure slightly. Let's create a single directory test_event_files and in that lets create the temp directories and files. Let me know what you think.
< jeffin143[m]>
walragatver: ok sounds ok
< jeffin143[m]>
But can you tell me the reason behind that option
< jeffin143[m]>
Like when do a user would turn it off , like he doesn't want the event files to be destroyed
< jeffin143[m]>
Also yesterday I was wrong
< jeffin143[m]>
Post command would run as soon as the target is build
< jeffin143[m]>
But we want to clear the dir everytime we run test
< jeffin143[m]>
And I don't think that is possible using cmake
< jeffin143[m]>
Cc @brim @shrit:matrix.org , let me know that is it possible to remove files using cmake after every run of test binary
< jeffin143[m]>
I don't think that's possible
ImQ009 has joined #mlpack
togo has joined #mlpack
< kartikdutt18Gitt>
@jeffin143, I might not have the right context, but If boost is there as a dependency, you can use boost::filesystem to remove files in runtime (after test / function call the remove function). And you can use ListDir to list all files and find the files you want to delete and only remove those.
< walragatver[m]>
jeffin143: While writing test you would be checking that event which you have written it's content is being displayed in tensorboard ui correctly right? So in order to view the content you would need those eventfiles after running test right?
< jeffin143[m]>
> @jeffin143, I might not have the right context, but If boost is there as a dependency, you can use boost::filesystem to remove files in runtime (after test / function call the remove function). And you can use ListDir to list all files and find the files you want to delete and only remove those.
< jeffin143[m]>
Currently I am not using boost in that and want to avoid it
< jeffin143[m]>
> jeffin143: While writing test you would be checking that event which you have written it's content is being displayed in tensorboard ui correctly right? So in order to view the content you would need those eventfiles after running test right?
< jeffin143[m]>
This would be manual correct ??
< kartikdutt18[m]>
> Currently I am not using boost in that and want to avoid it
< kartikdutt18[m]>
You can use std::remove as well then.
< jeffin143[m]>
That comes in c++ 17 @kartikdutt18:matrix.org
< jeffin143[m]>
I am not sure
< jeffin143[m]>
I will take a quick look
< kartikdutt18[m]>
Ahh.
< jeffin143[m]>
Yes that's since c++ 17
ghanima has joined #mlpack
favre49 has quit [Ping timeout: 260 seconds]
favre49 has joined #mlpack
< jeffin143[m]>
walragatver: I don't currently have a system to check whether tensorflow ui shows the summary properly or not
favre49 has quit [Ping timeout: 256 seconds]
favre49 has joined #mlpack
cult- has joined #mlpack
< cult->
hello, is the input sequence length into viterbi arbitrary or is there an optimal length for prediction?
< chopper_inbound[>
m.lozhnikov: I was trying to implement BLEU score metric. I've done the raw implementation https://gist.github.com/mrityunjay-tripathi/18fa5f09e02b422c7347c6462874a096 As I remember, n-gram is already implemented in mlpack. It would be nice if we can use it in this implementation...but I don't know how? I just wanted to simplify this script.
< mlozhnikov[m]>
chopper_inbound: If I remember right it's not merged yet. I did implemented it, but I didn't open a PR.
< chopper_inbound[>
m.lozhnikov: Oh...ok then. I will try to polish this up and open a pull request.
togo has quit [Read error: Connection reset by peer]
togo_ has quit [Ping timeout: 272 seconds]
< HimanshuPathakGi>
Hey @saksham189 I am trying to implement smo for multi-class classification but this implementation contains for two-class classification can you help me figure out. How can I calculate probabilities of multi class in smo.