turi::object_detection::Config Struct Reference

#include <toolkits/object_detection/od_model_trainer.hpp>

Public Attributes
int	max_iterations = -1
int	batch_size = -1
int	output_height = 13
int	output_width = 13
int	num_classes = -1

Detailed Description

Ostensibly model-agnostic parameters for object detection.

Definition at line 108 of file od_model_trainer.hpp.

Member Data Documentation

batch_size

int turi::object_detection::Config::batch_size = -1

The number of images to process per training batch.

If -1, then this size should be computed automatically.

Definition at line 121 of file od_model_trainer.hpp.

max_iterations

int turi::object_detection::Config::max_iterations = -1

The target number of training iterations to perform.

If -1, then this target should be computed heuristically.

Definition at line 114 of file od_model_trainer.hpp.

num_classes

int turi::object_detection::Config::num_classes = -1

Determines the number of feature channels in the final feature map.

Definition at line 130 of file od_model_trainer.hpp.

output_height

int turi::object_detection::Config::output_height = 13

For darknet-yolo, the height of the final feature map.

Definition at line 124 of file od_model_trainer.hpp.

output_width

int turi::object_detection::Config::output_width = 13

For darknet-yolo, the width of the final feature map.

Definition at line 127 of file od_model_trainer.hpp.

---

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

• toolkits/object_detection/od_model_trainer.hpp