Grasshopper Docs

Community documentation for Grasshopper add-ons & plugins

Evolutionary Learning
Explicit Components
Loop
Octopus
Supervised Learning
Utilities
View

Settings for RPROP Supervised Learning of an ANN

Inputs

NameIDDescriptionType
MaxStepsMaxStepsMaximum Steps for Learning; when reached, Learning ends. 0 for unlimited steps.Integer
MaxTimeMaxTimeTime Limit in milliseconds; 0 for no limit; when reached, Learning ends.Integer
MinErrMinErrMinimum Error between Input and Output for Learning; when reached, Learning ends.Number
Memory LimitMemLPercent of main memory to remain free; if dropping below, learning is stopped.Number
Divergence StepsDivSNumber of consecutive steps with worsening error after which learning is stopped. 0 to disable this check.Integer
HistHistRecord History? Set to true to get an ANN for each Step of the LearningBoolean
TypeTypeType of Resilient Propagation Method [0-RPROPp | 1-RPROPm | 2-iRPROPp | 3-iRPROPm]Integer
InitUpdInitUpdInitial update Parameter for Resilient Propagation AlgorithmNumber
MaxStepMaxStepMaxStep Parameter for Resilient Propagation AlgorithmInteger
CoresCoresNumber of processor cores to use; 0 to figure out the maximal optimum automatically.Integer
ResetResetSwitch to false to store the training model between updates of the component and continue training on it. The default resets the model on every update. If a SeedANN is supplied, training always starts from that one - then this switch does not affect anything.Boolean

Outputs

NameIDDescriptionType
SSSettings for Supervised Example Learning of Networks using Resilient Propagation from the Encog libraryRPROP Learning Settings

Site design © Robin Rodricks.   Octopus and associated data © 2020 Robert Vierlinger.  
Rhinoceros and Grasshopper are registered trademarks of Robert McNeel & Associates.  Hosted by GitHub

Report an Issue  |  Terms of Service