Model Class#

custom_neural_net_creator.model#

class custom_neural_net_creator.model.Model#

Bases: object

add(layer: Dense | ActivationLayer) → None#

Adds a layer to neural network model.

This method appends a layer, which can be either a Dense layer or Activation Layer, to the list of layers in the neural network model.

Parameters:

layer (Union[Dense, ActivationLayer]) – Layer to Be Added to Neural Network

fit(x: List, y: List, loss: mean_squared_error, loss_derivative: mean_squared_error_derivative, epochs: int | None = 100, learning_rate: float | None = 0.01, verbosity: int | None = 1)#

Trains the neural network model on the given data.

This method trains the neural network model using the specified input data x, target data y, loss function loss, and its derivative loss_derivative. It performs a number of training epochs specified by the epochs parameter, updating the model’s weights and biases through backpropagation with gradient descent.

The learning_rate parameter controls the step size during weight updates. The verbosity parameter determines the level of output during training, with options to print loss values at different intervals.

Parameters:

  • x (List) – Input data for training.

  • y (List) – Target data for training.

  • loss (mean_squared_error) – The loss function used for training.

  • loss_derivative (mean_squared_error_derivative) – The derivative of the loss function.

  • epochs (Optional[int], optional) – Number of training epochs. Defaults to 100.

  • learning_rate (Optional[float], optional) – Learning rate for gradient descent. Defaults to 0.01.

  • verbosity (Optional[int], optional) – Verbosity level for printing training progress. 0 - No output, 1 - Print every 100 epochs, 2 - Print every 10 epochs. Defaults to 1.

predict(input: List) → List#

Predicts the output for a given input using the neural network model.

This method takes a list of input data and computes predictions using the neural network model. It iterates through each input, performing forward propagation through the network’s layers and collecting the output. The predictions are stored in a list and returned.

Parameters:

input (List) – Input data to make predictions on.

Returns:

List of predictions for the input data.

Return type:

List

print_error(epoch: int, error: float) → None#

Prints the error for a given epoch.

This method prints the loss (error) for a specific training epoch. It is typically used for monitoring the training progress. The epoch parameter specifies the current epoch number, and error is the calculated loss for that epoch.

Parameters:

  • epoch (int) – The epoch number.

  • error (float) – The error value for the epoch.