synicix_ml_pipeline.models package

Submodules

synicix_ml_pipeline.models.MEImodel module

synicix_ml_pipeline.models.SimpleMLP module

class synicix_ml_pipeline.models.SimpleMLP.SimpleMLP(input_shape, output_shape, num_hidden_layers=1, hidden_size=1000, nonlinear_function_module_name='torch.nn', nonlinear_function_class_name='ELU', nonlinear_function_class_params={}, final_relu=False, l1_loss_lamda=0.0, l2_loss_lamda=0.0)[source]

Bases: torch.nn.modules.module.Module

A very simple example of a multilayer perceptron network illustrating the typical model design for synicix_ml_pipeline

compute_regularizer()[source]

Regularlization Computation Function

Parameters:
None
Returns:
Tensor: Regularlization loss to be added during the loss backpropergation
forward(x)[source]

Forward function for SimpleMLP

Parameters:
x (Tensor or list of Tensor): input batch, by default it only uses the first input tensor
Returns:
Tensor: Output of the model given the input Tensor: Regularlization loss to be added during the loss backpropergation

Module contents