java - DeepLearning4J IllegalArgumentException for CNN with custom Matrix -


i have custom 7(height) , 24(width) matrix input used training. output labels age ( young, mature, old). go deeplearning4j convolutional neural networks.

after building basic convolutional neural network first training item gives following error , have no clue about.

exception in thread "main" java.lang.illegalargumentexception: invalid size index 2 wher it's >= rank 2 @ org.nd4j.linalg.api.ndarray.basendarray.size(basendarray.java:4066) @ org.deeplearning4j.nn.layers.convolution.convolutionlayer.preoutput(convolutionlayer.java:192) @ org.deeplearning4j.nn.layers.convolution.convolutionlayer.activate(convolutionlayer.java:247) @ org.deeplearning4j.nn.graph.vertex.impl.layervertex.doforward(layervertex.java:88) @ org.deeplearning4j.nn.graph.computationgraph.feedforward(computationgraph.java:983) @ org.deeplearning4j.nn.graph.computationgraph.computegradientandscore(computationgraph.java:889)

my dl4j code 

//model config here multilayerconfiguration.builder builder = new neuralnetconfiguration.builder()     .seed(seed)     .iterations(iterations)     .regularization(true).l2(0.0005)     .learningrate(0.01)//.biaslearningrate(0.02)     //.learningratedecaypolicy(learningratepolicy.inverse).lrpolicydecayrate(0.001).lrpolicypower(0.75)     .weightinit(weightinit.xavier)     .optimizationalgo(optimizationalgorithm.stochastic_gradient_descent)     .updater(updater.nesterovs).momentum(0.9)      .list()     .layer(0, new convolutionlayer.builder(4, 1)         //nin , nout specify depth. nin here nchannels , nout number of filters applied             .name("hzvt1")         .nin(nchannels)         .stride(1, 1)         .nout(26)         .activation("relu")//.activation("identity")         .build())     .layer(1, new outputlayer.builder(lossfunctions.lossfunction.negativeloglikelihood)         .nout(outputnum)         .activation("softmax")         .build())       .setinputtype(inputtype.convolutional(nchannels,height,width))     .backprop(true).pretrain(false);  //model build here             model.fit(wmtrain);multilayerconfiguration conf = builder.build(); model.fit(wmtrain);multilayernetwork model = new multilayernetwork(conf); model.init();              //training data creation here  indarray weekmatrix = nd4j.ones(dlagegender.nchannels,dlagegender.height*dlagegender.width);        double[] vector = new double[] { 0.0, 1.0, 0.0 }; indarray intlabels = nd4j.create(vector); dataset ds=new dataset(weekmatrix,intlabels); //train first item model.fit(wmtrain);

i using dl4j version 0.6 , java version 1.8, maven 3.3+

i suspect bug in library.

with of gitter support. found out model , input not matching. correct working code follows.

i hope dl4j error/exception messages more clear in next releases.

log.info("build model...."); system.out.println("building model..."); multilayerconfiguration.builder builder = new neuralnetconfiguration.builder()         .seed(seed)         .iterations(iterations)         .regularization(true).l2(0.0005)         .learningrate(0.01)//.biaslearningrate(0.02)         //.learningratedecaypolicy(learningratepolicy.inverse).lrpolicydecayrate(0.001).lrpolicypower(0.75)         .weightinit(weightinit.xavier)         .optimizationalgo(optimizationalgorithm.stochastic_gradient_descent)         .updater(updater.nesterovs).momentum(0.9)         .list()         .layer(0, new convolutionlayer.builder(4, 1)             //nin , nout specify depth. nin here nchannels , nout number of filters applied             .name("hzvt1")             .nin(nchannels)             .stride(1, 1)             .nout(26)             .activation("relu")//.activation("identity")             .build())         .layer(1, new outputlayer.builder(lossfunctions.lossfunction.negativeloglikelihood)             .nout(classes)             .activation("softmax")             .build())         .setinputtype(inputtype.convolutional(height,width,nchannels))         .backprop(true).pretrain(false);  //model build here             model.fit(wmtrain);multilayerconfiguration conf = builder.build(); model.fit(wmtrain);multilayernetwork model = new multilayernetwork(conf); model.init();              //training data creation here      indarray weekmatrix = nd4j.ones(new int[]{1,dlagegender.nchannels,dlagegender.height,dlagegender.width});     indarray intlabels;     double[] vector = new double[] { 0.0, 1.0, };     intlabels = nd4j.create(vector); dataset ds=new dataset(weekmatrix,intlabels);  log.info("train model...."); model.setlisteners(new scoreiterationlistener(1)); model.fit(wmtrain); system.out.println("data train ok.");

Comments

Post a Comment

Popular posts from this blog

php - How to add and update images or image url in Volusion using Volusion API -

actually have store importing products store product data insert , update image or image url not insert , update. this xml code name datapro.txt. <products_joined> <productcode>3710_012t</productcode> <vendor_partno>eah5450silent/di/1gd3(lp)</vendor_partno> <productname>test product ta</productname> <hideproduct>n</hideproduct> <stockstatus>20</stockstatus> <lastmodified>1/5/2016 10:25:00 am</lastmodified> <lastmodby>2</lastmodby> <productweight>0.9</productweight> <productprice>100</productprice> <productmanufacturer>asus tek</productmanufacturer> <vendor_price>32.69</vendor_price> <numproductssharingstock>0</numproductssharingstock> <categoryids>107</categoryids> <producturl>http://tebkq.mvlce.servertrust.com/productdetails.asp?productcode=3710_012t</producturl>
Read more

javascript - jQuery UI Splitter/Resizable for unlimited amount of columns -

using jqueryui splitter or resizable functions want create layout allows unlimited amount of columns added , columns need resizable. so fiddle: https://jsfiddle.net/f3glh3mw/ but in fiddle html contains out of nested sets instead of inline. limits possibilities above fiddle can have equal widths 2,4,8,16,32 etc columns. in between shows crooked can see in fiddle. when researching came on question: jquery ui , splitter if @ top answer shows structure like: <div class="wrap"> <div class="resizable resizable1"></div> <div class="resizable resizable2"></div> </div> full fiddle: http://jsfiddle.net/8qztj/86/ which structure want javascript hardcoded 2 resizable items. so here question: how modify last fiddle works amount of .resizable divs. or know of plugin can this. thanks helping! so far found plugin: http://www.bacubacu.com/colresizable/ which says , close functionality want. in
Read more

javascript - IE9 error '$'is not defined -

Image
<!--[if lte ie 9]><html class="lte9"><![endif]--><!--[if gt ie 9]><html><![endif]--> <!--[if !ie]><!--> <html> <!--<![endif]--> blah blah <script src="https://ajax.googleapis.com/ajax/libs/jquery/1.12.1/jquery.min.js"></script> <script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.7/js/bootstrap.min.js"></script> <script src="{{ asset('js/mypage.js') }}" type="text/javascript"></script> </body> </html> when load page, ie9 brings error '$' not defined . how can solve error? +++ mypage.js $(function(){ //when load page, //ie9 tag not moving. $(window).load(function() { $('div.mask').hide(); }); // when click , willshow show. //for not ie 9 if (!$('html').hasclass('lte9')) { var = $('ul.about > li > span.activebtn');
Read more