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
<zoq[m]1> <turska79turska79> "> <@turska79-5fa54a47d73408ce4ff..." <- Right now the `-1` works because we use the empty loss function in the test case which doesn't interpret what the target type is. If you use the `MeanSquaredError` loss we have to make sure https://github.com/mlpack/mlpack/blob/master/src/mlpack/methods/ann/loss_functions/mean_squared_error_impl.hpp#L44 does return the right thing. Can you post the actual error you get?
<turska79turska79> from the line you mention compiler complains that: left of '.n_cols' must have class/struct/union.
<turska79turska79> As target intepreted as is int from -1 and it doesnt have member n_cols
Gulshan[m] has quit [Quit: Client limit exceeded: 20000]
KshitijAggarwal[ has quit [Quit: Client limit exceeded: 20000]
AbdullahKhilji[m has quit [Quit: Client limit exceeded: 20000]
RudraPatil[m] has quit [Quit: Client limit exceeded: 20000]
Prometheus[m] has quit [Quit: Client limit exceeded: 20000]
GauravTirodkar[m has quit [Quit: Client limit exceeded: 20000]
<turska79turska79> when template code is generated
VarunGupta[m] has quit [Quit: Client limit exceeded: 20000]
MohomedShalik[m] has quit [Quit: Client limit exceeded: 20000]
SoumyadipSarkar[ has quit [Quit: Client limit exceeded: 20000]
ShivamShaurya[m] has quit [Quit: Client limit exceeded: 20000]
ChaithanyaNaik[m has quit [Quit: Client limit exceeded: 20000]
RishabhGoel[m] has quit [Quit: Client limit exceeded: 20000]
ArunavShandeelya has quit [Quit: Client limit exceeded: 20000]
GauravGhati[m] has quit [Quit: Client limit exceeded: 20000]
HimanshuPathak[m has quit [Quit: Client limit exceeded: 20000]
ryan[m]1 has quit [Quit: Client limit exceeded: 20000]
<zoq[m]1> <turska79turska79> "> <@turska79-5fa54a47d73408ce4ff..." <- But looking at the code you posted, you converted modified it to use `rowvec` instead of an `int` for the target call type.
<turska79turska79> Ah yes
<turska79turska79> error before changing it to rowvec was that: left of '.n_cols' must have class/struct/union.
<turska79turska79> but im not sure what rowvec(-1) is compared to -1
<turska79turska79> if its array with one -1 value, then everything should be okay
<zoq[m]1> <turska79turska79> "but im not sure what rowvec(-1..." <- `arma::rowvec(-1)` does not work it should be `arma::rowvec("-1")` but even then I'm not sure what you are trying to do, because as you mentioned comparing with -1 makes not really sense, so you would have to generate your own target vecttor which depends on the task you like to solve.
<turska79turska79> Im trying to use as semi pseudo-code
<turska79turska79> mlpack::rl::SAC<ml::ContinuousActionEnvironment, FeedForwardNetwork, FeedForwardNetwork, ens::AdamUpdate, ReplayMethod>;
<turska79turska79> using FeedForwardNetwork = mlpack::ann::FFN<mlpack::ann::MeanSquaredError<arma::mat, arma::rowvec>, mlpack::ann::RandomInitialization>;
<turska79turska79> And MeanSquaredError is giving me hard time from its Forward/Backward methods
<turska79turska79> as this TargetType& target gets generated as int. and default int type doesnt have n_cols member... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/891744e427901e2ce18080e64abc91ede40970bc)
<turska79turska79> and i attempted to tell the compiler that TargetType template is rowvec
<turska79turska79> although calling MeanSquaredError forward with prediction and -1 as estimated value doesnt seem right
CaCode has joined #mlpack
CaCode has quit [Quit: Leaving]