pulsatrix 1.1.0
Loading...
Searching...
No Matches
pulsatrix::SoftDecisionTreeOptions Struct Reference

SoftDecisionTree's settings. More...

#include <soft_decision_tree.hpp>

Public Attributes

int depth = 4
 Levels of inner nodes: 2^depth leaves. Between 1 and 16.
 
float beta = 1.0f
 The gates' inverse temperature β: larger is closer to a hard tree.
 
float penalty = 0.0f
 The balance penalty's λ; 0 turns it off. Frosst and Hinton use 0.1 or so.
 
uint64_t seed = 0
 The gates' weights and biases start uniform in ±1/√inputs, as nn.Linear's do; the leaves start at zero, the uniform distribution.
 

Detailed Description

SoftDecisionTree's settings.

Member Data Documentation

◆ beta

float pulsatrix::SoftDecisionTreeOptions::beta = 1.0f

The gates' inverse temperature β: larger is closer to a hard tree.

◆ depth

int pulsatrix::SoftDecisionTreeOptions::depth = 4

Levels of inner nodes: 2^depth leaves. Between 1 and 16.

◆ penalty

float pulsatrix::SoftDecisionTreeOptions::penalty = 0.0f

The balance penalty's λ; 0 turns it off. Frosst and Hinton use 0.1 or so.

◆ seed

uint64_t pulsatrix::SoftDecisionTreeOptions::seed = 0

The gates' weights and biases start uniform in ±1/√inputs, as nn.Linear's do; the leaves start at zero, the uniform distribution.


The documentation for this struct was generated from the following file: