Available models

AlexNet

[source]

AlexNet

Dl4j’s AlexNet model interpretation based on the original paper ImageNet Classification with Deep Convolutional Neural Networksand the imagenetExample code referenced.References:http://papers.nips.cc/paper/4824-imagenet-classification-with-deep-convolutional-neural-networks.pdfhttps://github.com/BVLC/caffe/blob/master/models/bvlc_alexnet/train_val.prototxt

Model is built in dl4j based on available functionality and notes indicate where there are gaps waiting for enhancements.

Bias initialization in the paper is 1 in certain layers but 0.1 in the imagenetExample codeWeight distribution uses 0.1 std for all layers in the paper but 0.005 in the dense layers in the imagenetExample code

Darknet19

[source]

Darknet19Reference: https://arxiv.org/pdf/1612.08242.pdfImageNet weights for this model are available and have been converted from https://pjreddie.com/darknet/imagenet/using https://github.com/allanzelener/YAD2K .

There are 2 pretrained models, one for 224x224 images and one fine-tuned for 448x448 images.Call setInputShape() with either {3, 224, 224} or {3, 448, 448} before initialization.The channels of the input images need to be in RGB order (not BGR), with values normalized within [0, 1].The output labels are as per https://github.com/pjreddie/darknet/blob/master/data/imagenet.shortnames.list .

FaceNetNN4Small2

[source]

A variant of the original FaceNet model that relies on embeddings and triplet loss.Reference: https://arxiv.org/abs/1503.03832Also based on the OpenFace implementation: http://reports-archive.adm.cs.cmu.edu/anon/2016/CMU-CS-16-118.pdf

InceptionResNetV1

[source]

A variant of the original FaceNet model that relies on embeddings and triplet loss.Reference: https://arxiv.org/abs/1503.03832Also based on the OpenFace implementation: http://reports-archive.adm.cs.cmu.edu/anon/2016/CMU-CS-16-118.pdf

LeNet

[source]

LeNet was an early promising achiever on the ImageNet dataset.References:

• http://yann.lecun.com/exdb/publis/pdf/lecun-98.pdf
• https://github.com/BVLC/caffe/blob/master/examples/mnist/lenet.prototxtMNIST weights for this model are available and have been converted from https://github.com/f00-/mnist-lenet-keras.

NASNet

[source]

U-Net

Implementation of NASNet-A in Deeplearning4j. NASNet refers to Neural Architecture Search Network, a family of modelsthat were designed automatically by learning the model architectures directly on the dataset of interest.

This implementation uses 1056 penultimate filters and an input shape of (3, 224, 224). You can change this.

Paper: https://arxiv.org/abs/1707.07012ImageNet weights for this model are available and have been converted from https://keras.io/applications/.

ResNet50

[source]

Residual networks for deep learning.

Paper: https://arxiv.org/abs/1512.03385ImageNet weights for this model are available and have been converted from https://keras.io/applications/</a&gt;.

SimpleCNN

[source]

A simple convolutional network for generic image classification.Reference: https://github.com/oarriaga/face_classification/

SqueezeNet

[source]

U-Net

An implementation of SqueezeNet. Touts similar accuracy to AlexNet with a fraction of the parameters.

Paper: https://arxiv.org/abs/1602.07360ImageNet weights for this model are available and have been converted from https://github.com/rcmalli/keras-squeezenet/.

TextGenerationLSTM

[source]

LSTM designed for text generation. Can be trained on a corpus of text. For this model, numClasses is

Architecture follows this implementation: https://github.com/fchollet/keras/blob/master/examples/lstm_text_generation.py

Walt Whitman weights are available for generating text from his works, adapted from https://github.com/craigomac/InfiniteMonkeys.

TinyYOLO

[source]

Tiny YOLOReference: https://arxiv.org/pdf/1612.08242.pdf

ImageNet+VOC weights for this model are available and have been converted from https://pjreddie.com/darknet/yolousing https://github.com/allanzelener/YAD2K and the following code.

String filename = “tiny-yolo-voc.h5”;ComputationGraph graph = KerasModelImport.importKerasModelAndWeights(filename, false);INDArray priors = Nd4j.create(priorBoxes);

FineTuneConfiguration fineTuneConf = new FineTuneConfiguration.Builder().seed(seed).iterations(iterations).optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).gradientNormalization(GradientNormalization.RenormalizeL2PerLayer).gradientNormalizationThreshold(1.0).updater(new Adam.Builder().learningRate(1e-3).build()).l2(0.00001).activation(Activation.IDENTITY).trainingWorkspaceMode(workspaceMode).inferenceWorkspaceMode(workspaceMode).build();

ComputationGraph model = new TransferLearning.GraphBuilder(graph).fineTuneConfiguration(fineTuneConf).addLayer(“outputs”,new Yolo2OutputLayer.Builder().boundingBoxPriors(priors).build(),“conv2d_9”).setOutputs(“outputs”).build();

System.out.println(model.summary(InputType.convolutional(416, 416, 3)));

ModelSerializer.writeModel(model, “tiny-yolo-voc_dl4j_inference.v1.zip”, false);}</pre>

The channels of the 416x416 input images need to be in RGB order (not BGR), with values normalized within [0, 1].

UNet

[source]

U-Net

An implementation of U-Net, a deep learning network for image segmentation in Deeplearning4j.The u-net is convolutional network architecture for fast and precise segmentation of images.Up to now it has outperformed the prior best method (a sliding-window convolutional network) on the ISBI challenge forsegmentation of neuronal structures in electron microscopic stacks.

Paper: https://arxiv.org/abs/1505.04597Weights are available for image segmentation trained on a synthetic dataset

VGG16

[source]

VGG-16, from Very Deep Convolutional Networks for Large-Scale Image Recognitionhttps://arxiv.org/abs/1409.1556Deep Face Recognitionhttp://www.robots.ox.ac.uk/~vgg/publications/2015/Parkhi15/parkhi15.pdf

ImageNet weights for this model are available and have been converted from https://github.com/fchollet/keras/tree/1.1.2/keras/applications.CIFAR-10 weights for this model are available and have been converted using “approach 2” from https://github.com/rajatvikramsingh/cifar10-vgg16.VGGFace weights for this model are available and have been converted from https://github.com/rcmalli/keras-vggface.

VGG19

[source]

VGG-19, from Very Deep Convolutional Networks for Large-Scale Image Recognitionhttps://arxiv.org/abs/1409.1556ImageNet weights for this model are available and have been converted from https://github.com/fchollet/keras/tree/1.1.2/keras/applications.

Xception

[source]

U-Net

An implementation of Xception in Deeplearning4j. A novel deep convolutional neural network architecture inspired byInception, where Inception modules have been replaced with depthwise separable convolutions.

Paper: https://arxiv.org/abs/1610.02357ImageNet weights for this model are available and have been converted from https://keras.io/applications/.

YOLO2

[source]

YOLOv2Reference: https://arxiv.org/pdf/1612.08242.pdf

ImageNet+COCO weights for this model are available and have been converted from https://pjreddie.com/darknet/yolousing https://github.com/allanzelener/YAD2K and the following code.

String filename = “yolo.h5”;KerasLayer.registerCustomLayer(“Lambda”, KerasSpaceToDepth.class);ComputationGraph graph = KerasModelImport.importKerasModelAndWeights(filename, false);INDArray priors = Nd4j.create(priorBoxes);

FineTuneConfiguration fineTuneConf = new FineTuneConfiguration.Builder().seed(seed).optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).gradientNormalization(GradientNormalization.RenormalizeL2PerLayer).gradientNormalizationThreshold(1.0).updater(new Adam.Builder().learningRate(1e-3).build()).l2(0.00001).activation(Activation.IDENTITY).trainingWorkspaceMode(workspaceMode).inferenceWorkspaceMode(workspaceMode).build();

ComputationGraph model = new TransferLearning.GraphBuilder(graph).fineTuneConfiguration(fineTuneConf).addLayer(“outputs”,new Yolo2OutputLayer.Builder().boundingBoxPriors(priors).build(),“conv2d_23”).setOutputs(“outputs”).build();

System.out.println(model.summary(InputType.convolutional(608, 608, 3)));

ModelSerializer.writeModel(model, “yolo2_dl4j_inference.v1.zip”, false);}</pre>

The channels of the 608x608 input images need to be in RGB order (not BGR), with values normalized within [0, 1].

pretrainedUrl

public String pretrainedUrl(PretrainedType pretrainedType) 

Default prior boxes for the model