We would like to build a community for Small Basic programmers of any age who like to code. Everyone from total beginner to guru is welcome. Click here to register and share your programming journey!


Thread Rating:
  • 1 Vote(s) - 5 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Neural Nettwork (ANN) Extension
#31
JR,

This is a bug, fixed and uploaded a new version available with Extension Manager.
Reply
#32
I Formatted Code : QDZM107.000
But I Don't Get This Could You Explain.

   

   

   
ZS
Reply
#33
I assume you are asking what the outputs are, since the code you posted is just my original with File commands back in?

The textwindow has various output - see the code.
The coloured boxes are the results of using the trained ANN, they are displayed in pairs, where the left colour is the original R,G,B and the right is H,S,L trained using ANN.  First column of pairs come from the training data - you don't usually test with your training data - the other columns are new random data.
The plot is the RMS error of the training with Epoch.

If you have specific questions I can try to answer, but mostly its in the code...
[-] The following 1 user Likes litdev's post:
  • z-s
Reply
#34
What About This.
https://litdev.uk/mybb/showthread.php?ti...955#pid955
ZS
Reply
#35
LitDev,

I tried loading this by using extension manager and it still failed. Tried again using the beta download and it failed. Are you sure that you have it loaded?

JR
Reply
#36
Im out just now, but remember its the ann extension not litdev.  I may have done it wrong will check later.
Reply
#37
LitDev.

Tried with the extension manager a couple of times. Still no go.

JR
Reply
#38
JR,

Can't see any reason why it doesn't work - uploaded again and this time tested by installing from EM - SmallBasicANN version 1.2.0.4.  My test code, basically same as yours, obviously saved in folder that contains the XOrANN.txt file.

Code:
name = "XOrANN"
name = NeuralNetwork.Load(Program.Directory+"\"+name+".txt")
TextWindow.WriteLine(NeuralNetwork.GetWeights(name))
Reply
#39
LitDev,

I verified that the code is 1.2.04. Still got the same error. Index was outside the bounds of the array. I even tried it with regular Small Basic and got the same error.

You know this is really tough. We have an extension with 0 documentation. At least I can't find anything.

If you have any suggestions let me know!

I downloaded dnspy and am trying to use it. The thing I don't see is where an array is used. I don't even know where to look.

I did use your code snippet and that worked OK The main error is still "Index was ouside the bounds of the array, but with this now:
at SmallBasicANN.ANN.FeedForward()
at SmallBasicANN.ANN.Train(String fn, Boolean fm)
at SmallBasicANN.NeuralNetwork.Train(Primitive Name, Primitive TrainingSet, Primitive FileMode)
at _SmallBasicProgram._Main()

I am still using your original program ( ZWWL934.000). Trying to load the neural network. So, if you use that you might be able to reproduce.

JR
Reply
#40
Hii JR
Everything works fine for me try this code I have reformatted by removing comments and work fine if there is any error then I think code and extension are working good.
Code : QDZM107.000
ZS
Reply


Forum Jump:


Users browsing this thread: 2 Guest(s)