PT21-22-Reseau-Neurones/sobek/network.py
2021-12-09 18:50:57 +01:00

41 lines
1.5 KiB
Python
Executable File

import numpy as np
class network:
def __init__(self, inputLayerSize, *layerSizes):
if type(inputLayerSize) != int:
raise TypeError("The input layer size must be an int!")
self.__weights = []
self.__inputLayerSize = inputLayerSize
oldLayerSize = inputLayerSize
for layerSize in layerSizes:
self.__weights.append( np.random.default_rng(42).random((oldLayerSize, layerSize)) )
oldLayerSize = layerSize
self.__biases = [[0]*layerSize for layerSize in layerSizes]
self.__weights = np.array(self.__weights, dtype=object)
self.__biases = np.array(self.__biases, dtype=object)
def __reLu(value):
return max(0, value)
def process(self, input):
if type(input) != np.ndarray:
raise TypeError("The input must be a vector!")
if input.size != self.__inputLayerSize:
raise ValueError("The input vector has the wrong size!")
if input.dtype != np.float64:
raise TypeError("The input vector must contain floats!")
for layerWeights, bias in zip(self.__weights, self.__biases):
input = np.matmul(input, layerWeights)
input = np.add(input, bias)
#reLu application
with np.nditer(input, op_flags=['readwrite']) as layer:
for neuron in layer:
neuron = network.__reLu(neuron)
return input
def train(self, inputs, results):