scnn.loss_functions

Loss functions for training neural networks by convex reformulation.

Notes

• We only support the squared loss for the time being.

class scnn.loss_functions.SquaredLoss

The squared-error loss function.

Given predictions \(f(X)\) and targets \(y\), this loss function has the form,

\[L(f(X), y) = \frac{1}{2} \|f(X) - y\|_2^2.\]