Module gradient 

Gradient calculation utilities for neural networks

This module provides utilities for calculating gradients in the context of neural network training, focusing on efficiency and numerical stability.

Functions

binary_crossentropy_gradient

Calculate the gradient of binary cross-entropy with respect to predictions

hessian

Calculate the Hessian matrix for a scalar-valued function

jacobian

Calculate the Jacobian matrix for a function that maps from R^n to R^m

mse_gradient

Calculate the gradient of mean squared error with respect to predictions

softmax_crossentropy_gradient

Calculate the gradient of softmax cross-entropy with respect to logits