A Python module to create simple multi-layer perceptron neural networks using Levenberg-Marquardt training
This package uses Python 3.x This package requires numpy which can be downloaded using the following commands
pip install numpy
To install and use this package simply run:
pip install --index-url https://test.pypi.org/simple/ LM
Then you can simply import it using:
import LM
- Creating a new neural network
structure = [1, 5, 1]
nn = LM.NN(structure)
The list structure
describes how many layers there are in the network and how many neurons are in each layer sequentially
- Training the neural network
nn.train_lm(train_X, train_Y)
- Predicting new input
nn.predict(test_X)