Skip to content

Is a simple python library to use the concept of deep learning in your project (is verrrrrrryyyyyy simple for shur)

License

Notifications You must be signed in to change notification settings

LeLaboDuGame/SimplePyAI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

17 Commits
 
 
 
 
 
 
 
 

Repository files navigation

THIS GIT IS NOt UPDATE ! SEE TH PACKAGE ON PYPI ! THANKS TOU !

Is a simple python library to use the concept of deep learning in your project (is verrrrrrryyyyyy simple for shur)

SIMPLEPYAI

Hey guy ! I'm 15 years and a frensh dev ! I present you my project of neural network ! Just for the credit : LeLaboDuGame on Twitch -> https://twitch.tv/LeLaboDuGame You can use this library on all your project !

Now it's possible to choose an activation function, to do that you can refer at the "sigmoid" function and the "derivation_sigmoid" function

The model use the sigmoïde activation function.

HOW TO USE:

ON FIRST:

Download all of this library: pickle,matplotlib,tqdm,numpy

EXAMPLE:

x = [[1], [0]]
y = [[0], [1]]
nn = Neural_Network(x, y, layers=[1, 64, 32, 64, 1], learning_rate=0.001, activation_function=sigmoid,
                    derivation=derivation_sigmoid, reload_last_session=False)
nn.train(10000, show=True)
print(nn.predict([[1], [0]]))  # predict values

TO START:

-To init you must send 'x' and 'y' list (or numpy array).
'x' is your input and 'y' is your output.

-You must put the 'layers' (default is [2, 3, 1] 2 is the input and 1 the output neurone).
The 'layers' is a list represent all of your layers with the number of your neurone example:

[4,                                     3,4,10,1,                                       100]
 ^                                          ^                                             ^ 
 |                                          |                                             |
(is for 4 neurone in the first layer)(some layers)(And the number of neurones in the output)

-You can set the 'learning_rate' more is small more he understands, but it will take more repetition in
training.

-'reload_last_session' (default: False) is to reload the last session (if is the first sessions set to False)

TO TRAIN:

-'n_iter' is the number of train repetition

-'show' (default: True) is to show how the graphique of your model

-'save' (default: False) is to save your model (you can reload your model with the 'reload_last_session'
in init.

Thank you very much to use this library !

About

Is a simple python library to use the concept of deep learning in your project (is verrrrrrryyyyyy simple for shur)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages