ai.djl.basicmodelzoo.cv.object_detection.ssd

Class SingleShotDetection

java.lang.Object

ai.djl.nn.AbstractBlock

ai.djl.basicmodelzoo.cv.object_detection.ssd.SingleShotDetection

All Implemented Interfaces:

ai.djl.nn.Block

public final class SingleShotDetection
extends ai.djl.nn.AbstractBlock

SingleShotDetection is an implementation of Block that implements a Single Shot Detection (SSD) model for object detection.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SingleShotDetection.Builder The Builder to construct a SingleShotDetection.

Field Summary

Fields inherited from class ai.djl.nn.AbstractBlock

children, inputNames, inputShapes, parameters, version

Method Summary

Modifier and Type	Method and Description
static SingleShotDetection.Builder	builder() Creates a builder to build a SingleShotDetection.
protected ai.djl.ndarray.NDList	forwardInternal(ai.djl.training.ParameterStore parameterStore, ai.djl.ndarray.NDList inputs, boolean training, ai.djl.util.PairList<java.lang.String,java.lang.Object> params)
static ai.djl.nn.convolutional.Conv2d	getAnchorPredictionBlock(int numAnchors) Creates a anchor prediction block used in an SSD.
static ai.djl.nn.convolutional.Conv2d	getClassPredictionBlock(int numAnchors, int numClasses) Creates a class prediction block used in an SSD.
static ai.djl.nn.SequentialBlock	getDownSamplingBlock(int numFilters) Creates a Block that reduces the size of a convolutional block by half.
ai.djl.ndarray.types.Shape[]	getOutputShapes(ai.djl.ndarray.types.Shape[] inputShapes)
void	initialize(ai.djl.ndarray.NDManager manager, ai.djl.ndarray.types.DataType dataType, ai.djl.ndarray.types.Shape... inputShapes)
void	loadMetadata(byte version, java.io.DataInputStream is)

Methods inherited from class ai.djl.nn.AbstractBlock

addChildBlock, addParameter, beforeInitialize, cast, clear, describeInput, forward, forward, forwardInternal, getChildren, getDirectParameters, getParameters, initializeChildBlocks, isInitialized, loadParameters, prepare, readInputShapes, saveInputShapes, saveMetadata, saveParameters, setInitializer, setInitializer, setInitializer, toString

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Methods inherited from interface ai.djl.nn.Block

forward, validateLayout

Method Detail

forwardInternal

protected ai.djl.ndarray.NDList forwardInternal(ai.djl.training.ParameterStore parameterStore,
                                                ai.djl.ndarray.NDList inputs,
                                                boolean training,
                                                ai.djl.util.PairList<java.lang.String,java.lang.Object> params)

Specified by:

forwardInternal in class ai.djl.nn.AbstractBlock

getOutputShapes

public ai.djl.ndarray.types.Shape[] getOutputShapes(ai.djl.ndarray.types.Shape[] inputShapes)

initialize

public void initialize(ai.djl.ndarray.NDManager manager,
                       ai.djl.ndarray.types.DataType dataType,
                       ai.djl.ndarray.types.Shape... inputShapes)

Specified by:

initialize in interface ai.djl.nn.Block

Overrides:

initialize in class ai.djl.nn.AbstractBlock

loadMetadata

public void loadMetadata(byte version,
                         java.io.DataInputStream is)
                  throws java.io.IOException,
                         ai.djl.MalformedModelException

Overrides:

loadMetadata in class ai.djl.nn.AbstractBlock

Throws:

java.io.IOException

ai.djl.MalformedModelException

getDownSamplingBlock

public static ai.djl.nn.SequentialBlock getDownSamplingBlock(int numFilters)

Creates a Block that reduces the size of a convolutional block by half.

Parameters:

numFilters - the number of filters

Returns:

a Block that reduces the size of a convolutional block by half

getClassPredictionBlock

public static ai.djl.nn.convolutional.Conv2d getClassPredictionBlock(int numAnchors,
                                                                     int numClasses)

Creates a class prediction block used in an SSD.

Parameters:

numAnchors - the number of anchors

numClasses - the number of classes

Returns:

a class prediction block used in an SSD

getAnchorPredictionBlock

public static ai.djl.nn.convolutional.Conv2d getAnchorPredictionBlock(int numAnchors)

Creates a anchor prediction block used in an SSD.

Parameters:

numAnchors - the number of anchors

Returns:

a anchor prediction block used in an SSD

builder

public static SingleShotDetection.Builder builder()

Creates a builder to build a SingleShotDetection.

Returns:

a new builder