Configuring Loss/Optimizers

Function Signature

def config(
    optimizer: str = 'gradient descent',
    loss: str = 'mean square error'
) -> None:

Parameters

• optimizer (str, default='gradient descent'): The optimizer to use during training. Currently, deeprai is in beta, so the only valid option for optimizer is 'gradient descent'.

• loss (str, default='mean square error'): The loss function to use during training. Valid options are 'mean square error', 'categorical cross entropy', and 'mean absolute error'.

Return Value

This function does not return anything. It modifies the deeprai.models.FeedForward instance by setting the optimizer and loss function.

Description

The config function sets the optimizer and loss function for the deeprai.models.FeedForward instance. While it is not necessary to call this function, if called, it will use the default values of 'gradient descent' for optimizer and 'mean square error' for loss function.

Currently, deeprai is in beta, so the only valid option for optimizer is 'gradient descent'. The loss parameter sets the loss function to use during training. Valid options are 'mean square error', 'categorical cross entropy', and 'mean absolute error'.

Examples

Here's an example of how to use the config function:

from deeprai.models import FeedForward

model = FeedForward()
model.add_dense(784)
model.add_dense(128, activation='relu')
model.add_dense(64, activation='relu')
model.add_dense(10, activation='softmax')
model.config(optimizer='gradient descent', loss='categorical cross entropy')

This code creates a FeedForward model with an input shape of (784,), adds three dense layers with ReLU and softmax activation functions, and sets the optimizer to 'gradient descent' and the loss function to 'categorical cross entropy'.