NeuralFit provides a Keras-like API to evolve neural networks in Python, allowing anyone to apply complex neuro-evolution algorithms.
pip install neuralfitimport neuralfit as nf
model = nf.Model(3, 1)
model.compile('alpha', loss='mse')
model.evolve(x, y, epochs=100)
model.predict(x_test, y_test)
Neuro-evolution is a relatively unexplored area of machine learning, and we believe this is mainly due to a lack of easy-to-use libraries. We want to show people the potential of neuro-evolution, hoping to accelerate its research and development.
Our library is easy to install and can be used similar to the familiar Keras API. Trained models can be exported to Keras.
Not only do we provide standard documentation, we also provide minimal working examples to ease implementation.
Next to providing algorithms from literature, we develop our own algorithms. These algorithms will be published once growing stale.