By Genius_um / MazeGroup - 2023
Version 0.0.1
LNPackager is a Python Module for make Neuronnal Nodes
and make Package for train they.
import lnpackager as lnp
# Put lnpackager.py in Python Libs files or in the same folder your program
And make a package instance, train it and save it :
Exemple :
package = lnp.LNPackage("package_file", "ctr", lnp.MultiplesNodes(), 5, 500) # Init Package Class
package.train() # Train Nodes
package.make() # Save Package File
Functions and Classes Docs :
-
Package class.
Learning types :
- ctr (Check To Result)
More your attemps and nodes numbers
is big, more is your waiting time.*
-
A node as a value and a marge value.
At a generation, this value change
for a random value between the marge.
-
At a generation, the value change
for a random value between the marge.
-
You can make multiples nodes
with this function and choose
there values and marges.
-
Training : you can log the
advencement of your train but
it's more low.
CTR Method Training :
- For attemps:
- For nodes:
-
Generate nodes values
-
Choose random operator
-
Result calculated with operator*
-
- For nodes:
- For attemps:
-
Make .lnp package file
Syntax :
l-type-<training type>;valid-result:<valid result>;attemps:<attemps>; (!<node initial value>;~<node marge>;)<...> [[!<node value generated>;#<random operator>;?<result>;]<... * Nodes numbers>]<...>