#include <ml/neural_net/compute_context.hpp>

Public Attributes
int	batch_size

int	num_features

int	prediction_window

int	num_classes

int	num_predictions_per_chunk

int	random_seed

bool	is_training

float_array_map	weights

Detailed Description

A struct to define all the parameters used to create the activity classifier model backend

Definition at line 23 of file compute_context.hpp.

Member Data Documentation

◆ batch_size

int turi::neural_net::ac_parameters::batch_size

Defines the batch size

Definition at line 25 of file compute_context.hpp.

◆ is_training

bool turi::neural_net::ac_parameters::is_training

Set to true, when the data is used for training.

Definition at line 45 of file compute_context.hpp.

◆ num_classes

int turi::neural_net::ac_parameters::num_classes

Defines the number of classes

Definition at line 36 of file compute_context.hpp.

◆ num_features

int turi::neural_net::ac_parameters::num_features

Defines the number of features in the data

Definition at line 28 of file compute_context.hpp.

◆ num_predictions_per_chunk

int turi::neural_net::ac_parameters::num_predictions_per_chunk

Each session is segmented into chunks of this many prediction windows.

Definition at line 39 of file compute_context.hpp.

◆ prediction_window

int turi::neural_net::ac_parameters::prediction_window

Each group of this many consecutive samples from the same session are assumed to have the same class label.

Definition at line 33 of file compute_context.hpp.

◆ random_seed

int turi::neural_net::ac_parameters::random_seed

Setting random seed makes results reproducible.

Definition at line 42 of file compute_context.hpp.

◆ weights

float_array_map turi::neural_net::ac_parameters::weights

Defines the weights of the network

Definition at line 48 of file compute_context.hpp.

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

ml/neural_net/compute_context.hpp

Public Attributes