PyAiNetwork 0.3.3

A simple Python library for creating neural networks

PyAiNetwork

PyAiNetork - this is a lightweight library to create small to medium-sized LM's.

---

Main functions

• Creating neural networks
• neural network training
• learning and counting through numpy tables

Installation

pip install PyAiNetwork

---

examples of use normal

from PyAiNetwork import Network

net = Network(
    4, #input 
    1, #layers
    1, #neorons on layer
    4 #output
    'gelu' #activition function
)

input = [0.2,0.4,0.6,0.8]

output = net.forward(input)
print(output)

for i in range(1000):
    net.train(
        input, #input
        input, #output
        0.01 #learing rate
    )

output = net.forward(input)
print(output)

examples of use profi

from PyAiNetwork import ProfNetwork

layers_neorons = [1]

net = ProfNetwork(
    4, #input 
    layers_neorons, #layers
    4, #output
    'gelu', #activition function
    'matrix' # math type (or every)
)

input = [0.2,0.4,0.6,0.8]

output = net.forward(input)
print(output)

for i in range(1000):
    net.train(
        input, #input
        input, #output
        0.01 #learing rate
    )

output = net.forward(input)
print(output)

net.save("save.json") #json file name

net2 = ProfNetwork(
    4, #input 
    layers_neorons, #layers
    4, #output
    'gelu', #activition function
    'matrix' # math type (or every)
)

net2.load("save.json") #json file name

output = net2.forward(input)
print(output)

GitHub

https://github.com/eyes-studio/PyAiNetwork

---