basic example

[elpashuto]

Hello, im trying to use the package but i didnt find a basic simple example
basically i built a game and i want to train the SharpNEAT on it

my inputs for the network are double[] input // around 30 different params
and the output is a simple bool[] output // only 1 key at the moment (jump)

can i get the most basic example?

another small question, lets assume that we achieved the level of training we want.
how can i serialize the model into a file inorder to load it and just use the well-train model

THANKS!!

[colgreen]

Hi,

I suggest looking at the Prey Capture task, here:

• https://github.com/colgreen/sharpneat-refactor/blob/main/src/SharpNeat.Tasks/PreyCapture/PreyCaptureEvaluator.cs

For an example of a basic/minimal app for evolving a neural net genome for a specific problem task, see the TestApp, here:

• https://github.com/colgreen/sharpneat-refactor/tree/main/src/Tests/TestApp1

E.g. in Program.cs you can switch the call to CreateNeatEvolutionAlgorithm_Binary11 to CreateNeatEvolutionAlgorithm_PreyCapture or your own equivalent code that is based on CreateNeatEvolutionAlgorithm_PreyCapture.

---

Saving and loading of evolved neural nets for later re-use is straightforward with these handy static helper methods:

Save a genome (e.g. the current best) to file, like so:

NeatGenomeSaver.Save(bestGenome, filepath);

Load the saved file like so:

NetFileModel netFileModel = NetFile.Load(filename);

To convert netFileModel to a neural net that can be used, do:

IBlackBox<double> neuralNet = NeuralNetConverter.ToNeuralNet(netFileModel);

Hope that helps.

[elpashuto]

Thanks!!‫ב-17 בדצמ׳ 2022, בשעה 19:55, ‏‏Colin Green ***@***.***> כתב/ה:‬ Hi, I suggest looking at the Prey Capture task, here: https://github.com/colgreen/sharpneat-refactor/blob/main/src/SharpNeat.Tasks/PreyCapture/PreyCaptureEvaluator.cs For an example of a basic/minimal app for evolving a neural net genome for a specific problem task, see the TestApp, here: https://github.com/colgreen/sharpneat-refactor/tree/main/src/Tests/TestApp1 E.g. in Program.cs you can switch the call to CreateNeatEvolutionAlgorithm_Binary11 to CreateNeatEvolutionAlgorithm_PreyCapture or your own equivalent code that is based on CreateNeatEvolutionAlgorithm_PreyCapture. Saving and loading of evolved neural nets for later re-use is straightforward with these handy static helper methods: Save a genome (e.g. the current best) to file, like so: NeatGenomeSaver.Save(bestGenome, filepath); Load the saved file like so: NetFileModel netFileModel = NetFile.Load(filename); To convert netFileModel to a neural net that can be used, do: IBlackBox<double> neuralNet = NeuralNetConverter.ToNeuralNet(netFileModel); Hope that helps. —Reply to this email directly, view it on GitHub, or unsubscribe.You are receiving this because you authored the thread.Message ID: ***@***.***>