adrianna 0.0.7.2

lib for development with machine learning

A powerfull lib, for development in machine learning

---

Install:

$ pip install adrianna -U

Examples:

Bases

Neuro V1 (binary)

from adrianna.neuro.base_v1 import NeuralNetwork
import numpy as np


# Exemplo de uso
if __name__ == "__main__":
    # Dados de entrada e saída
    X = np.array([[3, 1, 2], [1, 24, 5], [2, 42, 5], [2, 23, 3]])
    y = np.array([[1], [0], [1], [0]])
    
    # Criação e treinamento da rede neural
    input_size = X.shape[1]  # Número de colunas das listas
    hidden_size = 4  # Número de neurônios na camada oculta
    output_size = y.shape[1]  # Número de saídas
    
    neural_net = NeuralNetwork(input_size, hidden_size, output_size, learning_rate=0.1)
    neural_net.train(X, y, epochs=10000)
    
    # Fazendo previsões
    predictions = neural_net.predict(X)
    predictions = np.round(predictions).astype(int)
    print("Predictions:")
    print(predictions)

Results:

...
Epoch 8000, Loss: 0.2617957
Epoch 8100, Loss: 0.1830651
Epoch 8200, Loss: 0.1800935
Epoch 8300, Loss: 0.2585192
Epoch 8400, Loss: 0.3310709
Epoch 8500, Loss: 0.3215849
Epoch 8600, Loss: 0.1803035
Epoch 8700, Loss: 0.1802555
Epoch 8800, Loss: 0.1807692
Epoch 8900, Loss: 0.3312975
Epoch 9000, Loss: 0.1800601
Epoch 9100, Loss: 0.2642676
Epoch 9200, Loss: 0.1930688
Epoch 9300, Loss: 0.3279387
Epoch 9400, Loss: 0.1871483
Epoch 9500, Loss: 0.1809427
Epoch 9600, Loss: 0.2635577
Epoch 9700, Loss: 0.1855325
Epoch 9800, Loss: 0.1796770
Epoch 9900, Loss: 0.1800897

Predictions:
[[1]
 [0]
 [1]
 [1]]