NEURAL_SAVE is not called with PARAMETERS+RULES

Gamestudio Links

Zorro Links

M1 Oversampling
by Petra. 04/24/24 10:34

Data from CSV not parsed correctly
by EternallyCurious. 04/20/24 21:39

Scripts not found
by juergen_wue. 04/20/24 18:51

zorro 64bit command line support
by 7th_zorro. 04/20/24 10:06

folder management functions
by VoroneTZ. 04/17/24 06:52

AUM Magazine

Latest Screens

5 registered members (AndrewAMD, Ayumi, PeWi, Quad, VoroneTZ), 513 guests, and 6 spiders.

Key: Admin, Global Mod, Mod

Newest Members

Print Thread

Rate Thread

NEURAL_SAVE is not called with PARAMETERS+RULES #453044 07/06/15 05:27 07/06/15 05:27
Joined: Sep 2013 Posts: 504 California G GPEngine OP User
GPEngine OP User G Joined: Sep 2013 Posts: 504 California	http://zorro-trader.com/manual/en/training.htm says Quote: The following cases are possible: Rules depend on a parameter. ... Set the RULES and PARAMETERS flags at the same time. The training run will now repeat through many cycles. At first the parameter will be optimized and for every parameter step a new set of rules will be generated. ... I find that when combining PARAMETERS with RULES with an adviseLong(NEURAL, ...), neural(mode=NEURAL_SAVE, ...) is not called as expected at the end of each optimize step. But, neural(model=NEURAL_LOAD,.. ) is called. Therefore Zorro dies when it tries to load an non-existing ml file since it wasn't produced, or it happily re-uses one from some old experiment.

Re: NEURAL_SAVE is not called with PARAMETERS+RULES [Re: GPEngine] #453045 07/06/15 05:38 07/06/15 05:38
Joined: Sep 2013 Posts: 504 California G GPEngine OP User
GPEngine OP User G Joined: Sep 2013 Posts: 504 California	A work-around is to use my own naming scheme for the ml files and save the model as part of NEURAL_TRAIN. This way, the right file is present when NEURAL_LOAD happens.

Re: NEURAL_SAVE is not called with PARAMETERS+RULES [Re: GPEngine] #453046 07/06/15 06:12 07/06/15 06:12
Joined: Sep 2013 Posts: 504 California G GPEngine OP User
GPEngine OP User G Joined: Sep 2013 Posts: 504 California	workaround fails because 'StepCycle' is undeclared identifier in 1.32. I need that to avoid naming collisions.

Re: NEURAL_SAVE is not called with PARAMETERS+RULES [Re: GPEngine] #453054 07/06/15 11:09 07/06/15 11:09
Joined: Jul 2000 Posts: 27,982 Frankfurt jcl Chief Engineer
jcl Chief Engineer Joined: Jul 2000 Posts: 27,982 Frankfurt	WFOCycle is normally used for rule naming. Optimize steps are counted with g->nStepCycle. Thanks for the report, we will look into combining NEURAL with PARAMETERS.

Moderated by Petra