0 ×

001_​Using_​DeepLearning4J_​to_​classify_​MNIST_​Digits

Workflow

Classifying handwritten digits using KNIME, DL4J and a LeNet variant
The workflow downloads, uncompresses and preprocesses the orignal MNIST dataset from: http://yann.lecun.com/exdb/mnist/ The two "Normalize Images" wrapped metanodes use the KNIME Streaming functionality to convert the input files into KNIME image cells that can be used by the DL4J Learner and Predictor. The "LeNet" metanode (taken from the Node Repository) is a variant of the originally described LeNet convolutional neural network. The images and the DL4J model is then used by the Learner to train a model (saved using the DL4J Model Writer), which is then applied to the test set, which is finally scored. This Tutorial is described in the blog post https://www.knime.org/blog/MNIST-DL4J-Intro Required Installations: Tools: KNIME Analytics Platform 3.3.1 (or greater) on your machine OR KNIME Cloud Analytics Platform on Azure Cloud OR KNIME Cloud Analytics Platform on AWS Cloud; Python 2.7.x configured for use with KNIME Analytics Platform: https://www.knime.org/blog/how-to-setup-the-python-extension
deep learning GPU image classification digit recognition Le Net
Classifying handwritten digits using KNIME, DL4J and a LeNet variant The workflow downloads, uncompresses and preprocesses the orignal MNIST dataset from: http://yann.lecun.com/exdb/mnist/The two "Normalize Images" wrapped metanodes use the KNIME Streaming functionality to convert the input files into KNIME image cells that can be used by the DL4J Learner andPredictor.The "LeNet" metanode (taken from the Node Repository) is a variant of the originally described LeNet convolutional neural network.The images and the DL4J model is then used by the Learner to train a model (saved using the DL4J Model Writer), which is then applied to the test set, which is finally scored.This Tutorial is described in the blog post https://www.knime.com/blog/learning-deep-learningRequired Installations:Tools: KNIME Analytics Platform 3.3.1 (or greater) on your machine OR KNIME Cloud Analytics Platform on Azure Cloud OR KNIME Cloud Analytics Platform on AWS CloudPython 2.7.x configured for use with KNIME Analytics Platform: https://www.knime.org/blog/how-to-setup-the-python-extension Extensions:KNIME Deeplearning4J extension from KNIME Labs Extensions/KNIME Deeplearning4J Integration (64 bit only). KNIME Image Processing extension from the KNIME Community Contributions - Image Processing and AnalysisKNIME Image Processing - Deep Learning 4J IntegrationVernalis KNIME Nodes from KNIME Community Contributions - Cheminformatics KNIME File Handling Nodes and KNIME Python Integration from KNIME & Extensions If you are running KNIME Analytics Platform on your machine:KNIME Image Processing - Deep Learning 4J Integration from the Stable Community Contributions update site (note this update site must be manually enabled, see Addendum fordetails) Optionally, if you have GPUs: GPU support (see blog post for details) The single layer Neural Networkarchitecture The five layer Neural Network(LeNet) architecture Double-click tosee the networkarchitectureInspect the trainingset imagesTrain the LeNetnetwork model(training will take some time)Make the predictionGenerate theconfusion matrixMake the predictionGenerate theconfusion matrixTrain a network modelSave the trained modelfor reuseNode 152 Download datasetand convert to CSV LeNet DL4J ModelInitializer Image Viewer DL4J Feedforward Learner(Classification) DL4J Feedforward Predictor(Classification) Scorer DL4J Feedforward Predictor(Classification) Scorer Dense Layer DL4J Feedforward Learner(Classification) DL4J Model Writer Normalizeimages (test) Normalizeimages (train) Download datasetand convert to CSV Normalizeimages (train) Normalizeimages (test) Missing Value Classifying handwritten digits using KNIME, DL4J and a LeNet variant The workflow downloads, uncompresses and preprocesses the orignal MNIST dataset from: http://yann.lecun.com/exdb/mnist/The two "Normalize Images" wrapped metanodes use the KNIME Streaming functionality to convert the input files into KNIME image cells that can be used by the DL4J Learner andPredictor.The "LeNet" metanode (taken from the Node Repository) is a variant of the originally described LeNet convolutional neural network.The images and the DL4J model is then used by the Learner to train a model (saved using the DL4J Model Writer), which is then applied to the test set, which is finally scored.This Tutorial is described in the blog post https://www.knime.com/blog/learning-deep-learningRequired Installations:Tools: KNIME Analytics Platform 3.3.1 (or greater) on your machine OR KNIME Cloud Analytics Platform on Azure Cloud OR KNIME Cloud Analytics Platform on AWS CloudPython 2.7.x configured for use with KNIME Analytics Platform: https://www.knime.org/blog/how-to-setup-the-python-extension Extensions:KNIME Deeplearning4J extension from KNIME Labs Extensions/KNIME Deeplearning4J Integration (64 bit only). KNIME Image Processing extension from the KNIME Community Contributions - Image Processing and AnalysisKNIME Image Processing - Deep Learning 4J IntegrationVernalis KNIME Nodes from KNIME Community Contributions - Cheminformatics KNIME File Handling Nodes and KNIME Python Integration from KNIME & Extensions If you are running KNIME Analytics Platform on your machine:KNIME Image Processing - Deep Learning 4J Integration from the Stable Community Contributions update site (note this update site must be manually enabled, see Addendum fordetails) Optionally, if you have GPUs: GPU support (see blog post for details) The single layer Neural Networkarchitecture The five layer Neural Network(LeNet) architecture Double-click tosee the networkarchitectureInspect the trainingset imagesTrain the LeNetnetwork model(training will take some time)Make the predictionGenerate theconfusion matrixMake the predictionGenerate theconfusion matrixTrain a network modelSave the trained modelfor reuseNode 152Download datasetand convert to CSV LeNet DL4J ModelInitializer Image Viewer DL4J Feedforward Learner(Classification) DL4J Feedforward Predictor(Classification) Scorer DL4J Feedforward Predictor(Classification) Scorer Dense Layer DL4J Feedforward Learner(Classification) DL4J Model Writer Normalizeimages (test) Normalizeimages (train) Download datasetand convert to CSV Normalizeimages (train) Normalizeimages (test) Missing Value

Download

Get this workflow from the following link: Download

Resources

Nodes

001_​Using_​DeepLearning4J_​to_​classify_​MNIST_​Digits consists of the following 135 nodes(s):

Plugins

001_​Using_​DeepLearning4J_​to_​classify_​MNIST_​Digits contains nodes provided by the following 7 plugin(s):