Regularization

Module: generative_models.core.losses.regularization

Source: generative_models/core/losses/regularization.py

Overview

Regularization losses module.

This module provides regularization terms that can be added to other losses to improve model stability, generalization, and prevent overfitting. All functions are JAX-compatible and work with NNX modules.

Classes

DropoutRegularization

class DropoutRegularization

SpectralNormRegularization

class SpectralNormRegularization

Functions

call

def __call__()

call

def __call__()

init

def __init__()

init

def __init__()

disc_interpolated

def disc_interpolated()

exclude_bias_predicate

def exclude_bias_predicate()

exclude_norm_predicate

def exclude_norm_predicate()

gradient_penalty

def gradient_penalty()

l1_regularization

def l1_regularization()

l2_regularization

def l2_regularization()

only_conv_predicate

def only_conv_predicate()

orthogonal_regularization

def orthogonal_regularization()

spectral_norm_regularization

def spectral_norm_regularization()

total_variation_loss

def total_variation_loss()

Module Statistics

• Classes: 2
• Functions: 14
• Imports: 5