rcurtin_irc changed the topic of #mlpack to: mlpack: a scalable machine learning library (https://www.mlpack.org/) -- channel logs: https://libera.irclog.whitequark.org/mlpack -- NOTE: messages sent here might not be seen by bridged users on matrix, gitter, or slack
robobub has quit [Quit: Connection closed for inactivity]
lambert[m] has quit [*.net *.split]
SlackIntegration has quit [*.net *.split]
YashBhagwatBlitz has quit [*.net *.split]
ShubhamAgrawal[m has quit [*.net *.split]
vaibhavp[m] has quit [*.net *.split]
psydroid has quit [*.net *.split]
SlackIntegration has joined #mlpack
psydroid has joined #mlpack
vaibhavp[m] has joined #mlpack
ShubhamAgrawal[m has joined #mlpack
lambert[m] has joined #mlpack
YashBhagwatBlitz has joined #mlpack
SameerPatil[m] has joined #mlpack
_slack_mlpack_14 has quit [Quit: You have been kicked for being idle]
coatless[m] has quit [Quit: You have been kicked for being idle]
<IWNMWEIWNMWE[m]> Hey can some one guide me in solving this issue giving be static code analysis fail in my pr
<IWNMWEIWNMWE[m]> I did not change anything related to this test
m-indn-sedta[m] has joined #mlpack
<akhunti1[m]> Hi All ,
<akhunti1[m]> I am using classify method from Random forest class like this [ rf.Classify(arma::mat(res), testPredictions, probs); ] for inferencing .
<akhunti1[m]> when i print probs it is retuning me 2 Probabilities[ As it is a binary class classification problem when i took sum it is 1 ].
<akhunti1[m]> but for inferencing we need single out put . so for that i did indexing [0,0].
<akhunti1[m]> but what i observed is index[0,0] is not always return highest Probability among 2 Probabilities.
<akhunti1[m]> so, just wanted to check do you have any other way , where it will return always highest Probability among 2 Probabilities for inferencing ?
_slack_mlpack_25 has quit [Quit: You have been kicked for being idle]
_slack_mlpack_25 has joined #mlpack
<shayan823ShayanS> Hi everyone,
<shayan823ShayanS> I am Shayan Shafquat. I have done my majors in Mathematics and Computing. I am currently pursuing my interest at the intersection of neuroscience and AI. Going through the GSoC proposals (though late), I got interested in contributing through the RL problem statement. I was setting up the environment and facing some issues. I have installed armadillo perfectly, and installed ensmallen using cmake. When I am running a simple SGD module
<shayan823ShayanS> on the code. It throws the following [error](https://paste.centos.org/view/642e99a5). The script I am trying to compile is at the given [link](https://paste.centos.org/view/20910d4f). I have tried solving it. Please help me identify the issue. Thanks.
<vaibhavp[m]> <shayan823ShayanS> "I am Shayan Shafquat. I have..." <- The namespace for the ensmallen library is 'ens' not 'ensmallen'. You can try changing that and then see if any other error exists.
<vaibhavp[m]> shayan823 (Shayan Shafquat): You can look at this [example](https://github.com/mlpack/ensmallen/blob/master/example.cpp) for reference.
<zoq[m]> <IWNMWEIWNMWE[m]> "image.png" <- You can ignore the issue, it's a false report.
<zoq[m]> <shayan823ShayanS> "I am Shayan Shafquat. I have..." <- Is ensmallen in your include search path? Not sure how you installed ensmallen.
<vaibhavp[m]> rcurtin: zoq: So about the DAG Network class, I had the thought that why shouldn't it be called NN class because not only directed acyclic graphs but also directed cyclic graphs can be implemented using the graph based approach, which will be very useful for certain cases? This will unify the FFN and RNN classes and make it flexible to create models.
<zoq[m]> <vaibhavp[m]> "rcurtin: zoq: So about the DAG..." <- Eventually yes, right now FFN and RNN implement the training strategy which is architecture specific.
<vaibhavp[m]> I also think, the Layer class needs to have a Forward, Backward and Gradient which accepts a std::vector of mat as input and output. I think this would be the best way forward to accommodate for a general NN class.
lxi has joined #mlpack
lxi has quit [Ping timeout: 260 seconds]