Weight Decay

The model trained by this node minimizes a quadratic loss. Weight Decay is a coefficient that scales a sum of squared weights (excluding bias parameters) that is a summand of this loss. Increasing the Weight Decay may improve the generalization of the model, while sacrificing how closely it fits the example targets. Increasing the value of this parameter may also help you get a more stable solution in cases that are difficult to train.

Error Threshold

This is a relative error threshold that is used to detect cases where the solution is not accurate enough. If the error of the computed model exceeds this threshold, then the node will display an error. This threshold doesn’t really do anything except generating errors when it is exceeded. This is to avoid surprises when an accurate model cannot be computed.

Kernel Type

This determines the type of kernel function that the model is based on. The kernel maps each pair of input components to a value.

Gaussian

Exponential function of a scaled, negated squared distance of two inputs.

Polynomial

Takes the dot product of two inputs, adds a constant to it and then raises the result to a specified degree.

Sigmoid

Takes the dot product of two inputs, multiplies that by a constant, adds another constant, and then applies a hyperbolic tangent.

Width

When in Gaussian, missing description.

Polynomial Offset

When in Polynomial, missing description.

Polynomial Offset

When in Polynomial, missing description.

Sigmoid Scale

When in Sigmoid, missing description.

Sigmoid Offset

When in Sigmoid, missing description.

Geometry nodes