03:43 < jeffin143[m]> rcurtin: https://github.com/mlpack/mlpack/pull/2844#issuecomment-789245194

03:43 < jeffin143[m]> Until and unless you approve we can't merge

03:43 < jeffin143[m]> Since you wanted to change

07:42 < shrit[m]> @ryan, if you now the reason why the error is caused please let me know, I was able to produce the issue too, but I do not know where it is comming from. Thus, I thought it is related to tree issue

14:40 < shrit[m]> rcurtin: It seems that the dynamic linked binaries are passing the test

14:41 < shrit[m]> While the cross compiled statically linked one are not passing

14:42 < shrit[m]> Here is the output from GDB:

15:54 < rcurtin[m]1> sorry about that jeffin143 , done now :)

16:10 < rcurtin[m]1> shrit: that is very weird that the test passes fine when dynamically linking... it would be a lot of output, but maybe you can instrument the dual-tree recursion itself? I can try and find some time but it would likely be tomorrow. basically you could add some information about when `BaseCase()` is called and when `Score()` is called (e.g. you could print the query and reference indices for `BaseCase()`, and print the

16:10 < rcurtin[m]1> query node first point index and number of points as well as the same information for the reference node in `Score()`)

16:14 < shrit[m]> I will add more debugging information in both side and compare them and see what is happening

16:14 < rcurtin[m]1> yes, it is definitely worth seeing if the recursion is different when dynamically linked vs. statically linked

16:14 < rcurtin[m]1> (you might want to hardcode a random seed)

18:57 < turska79Gitter[m> I have been doing snake game with mlpack every now and then for learning purposes, but now im a bit stuck. I have based my stuff on bipedal walker example, but my snake doesnt seem to be learning or improvint itself at all. (Im not using any training data for it)

18:57 < turska79Gitter[m> Now it seems to be running up and to wall with few turns on every game. So my code must have issues somewhere either with environment or how i confgure and use network.

18:57 < turska79Gitter[m> So any hints and/or tips would be much appreciated

19:26 < turska79Gitter[m> relevant bits being in here https://github.com/turska79/ml-skynet-snake/tree/machine-learning/ml-skynet-snake/ml-skynet-snake/ml and in LearningAgent.cpp

19:50 < shrit[m]> The question is why the second tree (the smaller one) is failing in the statically built and not on the dynamic built one?

19:51 < shrit[m]> I will add the score, and BaseCase later, I am trying to understand what is happening with this output

19:51 < rcurtin[m]1> are the datasets exactly the same for both of those runs?

19:52 < rcurtin[m]1> (might want to print that just to double-check)

19:52 < shrit[m]> I did not provide dataset. in the commandline

19:53 < rcurtin[m]1> right, if it's randomly generated in the test, definitely we want to make sure it is the same data

19:53 < rcurtin[m]1> maybe it's worth just hardcoding it to use a few points from some randomly generated csv or something

19:54 < shrit[m]> I will try to add the same dataset, I will look at the test

23:29 < zoq> Okay, that was unexpected ...

23:30 < zoq> SN10 is gone, I guess.

23:45 < rcurtin[m]1> wow, impressive landing... and impressive boom

23:46 < zoq> The landing was cool.

23:48 < zoq> turska79Gitter[m: When I have a chance I'll take a look at the code.