debut perceptron multicouche

neuralnetworkbuilder/test.py

@@ -0,0 +1,49 @@
+import numpy as np
+
+class network:
+
+    def __init__(self, inputLayerSize, *layerSizes):
+        self.weights = []
+        self.inputLayerSize = inputLayerSize
+        self.oldLayerSize = inputLayerSize
+        for layerSize in layerSizes:
+            self.weights.append( np.random.default_rng(42).random((self.oldLayerSize, layerSize)) )
+            self.oldLayerSize = layerSize
+        self.biases = [[0]*layerSize for layerSize in layerSizes]
+        self.weights = np.array(self.weights)
+        self.biases = np.array(self.biases)
+
+    def reLu(value):
+        return max(0, value)
+
+    def process(self, input):
+        if type(input) != np.ndarray:
+            print("non")
+        if input.size != self.inputLayerSize:
+            print("vite")
+        if input.dtype != np.float64:
+            print("aaa")
+        for layer, bias in zip(self.weights, self.biases):
+            print("---------------------")
+            print(input)
+            print(layer)
+            print(bias)
+            input = np.matmul(input, layer)
+            input = np.add(input, bias)
+            with np.nditer(input, op_flags=['readwrite']) as layer:
+                for neuron in layer:
+                    neuron = network.reLu(neuron)
+        return input
+
+
+test = network(16, 16, 8, 4)
+
+for y in test.weights:
+    print(y, end="\n\n")
+
+for y in test.biases:
+    print(y, end="\n\n")
+
+print(network.reLu(8))
+
+print(test.process(np.random.default_rng(42).random((16))))