Darknet.Darknet Class Reference

This class allows you to work with Darknet/YOLO neural networks. More...

Public Member Functions
	Darknet ()
	Constructor.
void	fix_out_of_bound_values (boolean toggle) throws Throwable
	Fix out-of-bound values, meaning coordinates, widths, or heights that extend past the edges of images will be cropped or adjusted.
void	free_image (MemorySegment image) throws Throwable
	Free an image that was loaded with Darknet.Darknet.load_image().
void	free_neural_network () throws Throwable
	Free the Darknet/YOLO neural network loaded by Darknet.Darknet.load_neural_network().
MemorySegment	load_image (String filename) throws Throwable
	Load the given image and convert it to a "DarknetImage" type.
MemorySegment	load_neural_network (String config, String names, String weights) throws Throwable
	Load the Darknet/YOLO neural network.
Map< String, Integer >	network_dimensions () throws Throwable
	Get the network dimensions.
void	set_detection_threshold (float threshold) throws Throwable
	Set the detection threshold.
void	set_gpu_index (int idx) throws Throwable
	Set the zero-based GPU index Darknet should use.
void	set_non_maximal_suppression_threshold (float threshold) throws Throwable
	Set the NMS threshold.
void	set_output_stream (String output_filename) throws Throwable
	Set the filename into which Darknet will write console output.
void	set_trace (boolean flag) throws Throwable
	Toggle trace output.
void	set_verbose (boolean flag) throws Throwable
	Toggle verbose output.
void	show_version_info () throws Throwable
	Display version information on STDOUT.
void	testing () throws Throwable
String	version_short () throws Throwable
	Get the short version string.
String	version_string () throws Throwable
	Get the full version string.

Detailed Description

This class allows you to work with Darknet/YOLO neural networks.

Since

2025-09-09

Constructor & Destructor Documentation

Darknet()

Darknet.Darknet.Darknet

(

)

Constructor.

See also

Darknet.Darknet.load_neural_network()

Since

2025-09-09

Member Function Documentation

fix_out_of_bound_values()

void Darknet.Darknet.fix_out_of_bound_values

(

boolean

toggle

)

throws Throwable

Fix out-of-bound values, meaning coordinates, widths, or heights that extend past the edges of images will be cropped or adjusted.

Default setting is true.

Since

2025-09-09

free_image()

void Darknet.Darknet.free_image

(

MemorySegment

image

)

throws Throwable

Free an image that was loaded with Darknet.Darknet.load_image().

Since

2025-09-09

free_neural_network()

void Darknet.Darknet.free_neural_network

(

)

throws Throwable

Free the Darknet/YOLO neural network loaded by Darknet.Darknet.load_neural_network().

It is safe to call this multiple times.

Note

Every neural network loaded with Darknet.Darknet.load_neural_network() must be freed with Darknet.Darknet.free_neural_network().

Since

2025-09-09

load_image()

MemorySegment Darknet.Darknet.load_image

(

String

filename

)

throws Throwable

Load the given image and convert it to a "DarknetImage" type.

Note

Remember to call Darknet.Darknet.free_image() when done with the image.

Since

2025-09-09

load_neural_network()

MemorySegment Darknet.Darknet.load_neural_network	(	String	config,
		String	names,
		String	weights
	)		throws Throwable

Load the Darknet/YOLO neural network.

Many Darknet calls require this to have been called, since this sets the neural network pointer. Remember to call Darknet.Darknet.free_neural_network() when done.

Note

Every neural network loaded with Darknet.Darknet.load_neural_network() must be freed with Darknet.Darknet.free_neural_network().

Since

2025-09-09

network_dimensions()

Map< String, Integer > Darknet.Darknet.network_dimensions

(

)

throws Throwable

Get the network dimensions.

The map will contain 3 entries: width, height, and channel.

Since

2025-09-09

set_detection_threshold()

void Darknet.Darknet.set_detection_threshold

(

float

threshold

)

throws Throwable

Set the detection threshold.

Default threshold setting is 0.25.

Since

2025-09-09

set_gpu_index()

void Darknet.Darknet.set_gpu_index

(

int

idx

)

throws Throwable

Set the zero-based GPU index Darknet should use.

• If you have a single GPU and Darknet/YOLO was built with support for your GPU, then the only valid value for the index will be 0 (zero) which is the default.

• If you have multiple GPUs and Darknet/YOLO was built with support for your GPUs, then you may want to select a specific GPU to use. If unspecified, then Darknet/YOLO will default to GPU index #0 (zero), meaning the first GPU.

• If Darknet/YOLO was built for CPU-only, then you should not call this API, or call it with -1.

Note

If called, this must be prior to Darknet.Darknet.load_neural_network().

Since

2025-09-09

set_non_maximal_suppression_threshold()

void Darknet.Darknet.set_non_maximal_suppression_threshold

(

float

threshold

)

throws Throwable

Set the NMS threshold.

Default NMS setting is 0.45.

Since

2025-09-09

set_output_stream()

void Darknet.Darknet.set_output_stream

(

String

output_filename

)

throws Throwable

Set the filename into which Darknet will write console output.

Default is STDOUT. This may be called prior to Darknet.Darknet.load_neural_network(). Use a blank filename ("") to reset it to the default STDOUT.

Since

2025-09-09

set_trace()

void Darknet.Darknet.set_trace

(

boolean

flag

)

throws Throwable

Toggle trace output.

This is only intended for debugging. Default trace setting is false. This may be called prior to Darknet.Darknet.load_neural_network().

Note

Turning on trace output will also turn on verbose. (You canot have trace without also having verbose.)

See also

Darknet.Darknet.set_verbose()

Darknet.Darknet.set_output_stream()

Since

2025-09-09

set_verbose()

void Darknet.Darknet.set_verbose

(

boolean

flag

)

throws Throwable

Toggle verbose output.

Default verbose setting is false. This may be called prior to Darknet.Darknet.load_neural_network().

Note

Turning off verbose output will also turn off trace (if it had been enabled).

See also

Darknet.Darknet.set_trace()

Darknet.Darknet.set_output_stream()

Since

2025-09-09

show_version_info()

void Darknet.Darknet.show_version_info

(

)

throws Throwable

Display version information on STDOUT.

This may be called prior to Darknet.Darknet.load_neural_network().

Example:

Darknet V5 "Moonlit" v5.0-138-ga061d2f0 [v5]
CUDA runtime version 12000 (v12.0), driver version 12060 (v12.6)
cuDNN version 12020 (v8.9.7), use of half-size floats is ENABLED
=> 0: NVIDIA GeForce RTX 3090 [#8.6], 23.6 GiB
Protobuf 3.21.12, OpenCV 4.6.0, Ubuntu 24.04

See also

Darknet.Darknet.version_string()

Darknet.Darknet.version_short()

Since

2025-09-09

testing()

void Darknet.Darknet.testing

(

)

throws Throwable

version_short()

String Darknet.Darknet.version_short

(

)

throws Throwable

Get the short version string.

For example, "5.0.138". This may be called prior to Darknet.Darknet.load_neural_network().

See also

Darknet.Darknet.version_string()

Since

2025-09-09

version_string()

String Darknet.Darknet.version_string

(

)

throws Throwable

Get the full version string.

For example, "v5.0-138-ga061d2f0". This may be called prior to Darknet.Darknet.load_neural_network().

See also

Darknet.Darknet.version_short()

Since

2025-09-09

---

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

• Darknet.java