ai.djl.basicmodelzoo.cv.classification

Class VGG

java.lang.Object

ai.djl.basicmodelzoo.cv.classification.VGG

public final class VGG
extends java.lang.Object

VGG model from the "Very Deep Convolutional Networks for Large-Scale Image Recognition" https://arxiv.org/abs/1409.1556 paper.

See Also:

The D2L chapter on VGG

Nested Class Summary

Nested Classes

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

Method Summary

Modifier and Type	Method and Description
static VGG.Builder	builder() Creates a builder to build a VGG.
static ai.djl.nn.Block	vgg(VGG.Builder builder) Creates a VGG block with the help of the VGG Builder.
ai.djl.nn.SequentialBlock	vggBlock(int numConvs, int numChannels) Creates a constituent VGG block that becomes a part of the whole VGG model.

Methods inherited from class java.lang.Object

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

Method Detail

vgg

public static ai.djl.nn.Block vgg(VGG.Builder builder)

Creates a VGG block with the help of the VGG Builder.

Parameters:

builder - the VGG.Builder with the necessary arguments

Returns:

a VGG block.

vggBlock

public ai.djl.nn.SequentialBlock vggBlock(int numConvs,
                                          int numChannels)

Creates a constituent VGG block that becomes a part of the whole VGG model.

Parameters:

numConvs - Numbers of layers in each feature block.

numChannels - Numbers of filters in each feature block. List length should match the layers.

Returns:

a constituent vgg block.

builder

public static VGG.Builder builder()

Creates a builder to build a VGG.

Returns:

a new builder