Set input array => Forward calculation of the individual => Get output array.

Returns the size of the best individual.

Multi threaded version of back propagation.

Single threaded version of back propagation.

Generate a random number in the interval [0, N) (uint32_t).

Adding the connections of neural network in preheating process.

Deep clone of individual.

Mutating the individuals.

Resize the capacity of the neural network cluster.

One of the functor of BP_Multi_Thread().

One of the functor of BP_Multi_Thread().

Generate a random number in the interval [min, max) (float).

Adding hidden nodes of neural network in preheating process.

BP_Single_Thread() is invoked in every individual of the neural network cluster.

The functor of backPropagation().

Check if the individualList is initialized.

All the individuals in the neural network cluster are sorted and the better ones are selected.

Turn off memory-limit.

A comparison functor used in sorting individuals.

Forward Computing of the best individual.

Parallel forward Computing of the best individual.

The functor of computeBatch().

Cross-Entropy.

Keep only the best one in the neural network cluster.

Euclidean distance.

Evolve the neural network cluster.

Fit all neural networks in the neural network cluster.

Fit all neural networks in the neural network cluster.

Returns capacity of Lycoris.

* Get a vector of nodes in a specific layer of the best * individual. The parameter pos starts at index 0. */

* Get a vector of nodes in a specific layer of this neural network. * The parameter pos starts at index 0. */

Returns the input dimension.

Returns the number of nodes in each layer of the neural network.

Returns the loss.

Returns mode of Lycoris (classify or predict).

Returns the output dimension.

Return the size of the individual.

Initialize a individual.

Import the pre-trained model.

Import the pre-trained model via string.

Mutating.

The functor of mutate().

Turn on memory-limit.

Overloading operator.

Preheating process of the neural network cluster.

The functor of preheat().

ReLU (with huge leak = 0.2).

Main part of evolve().

Export the current trained model.

Set the number of worker threads to train the model.

Set the learning rate.

Set p1 to p4 in the class Args.

Set the odds of mutating.

Softmax.

Return a list of the words in the string, using whitespace as the delimiter string.