christopher commited on
Commit
bcc5b6d
1 Parent(s): e27cf6d

Upload 7 files

Browse files
Files changed (7) hide show
  1. analytics.js +1 -1
  2. bundle.css +0 -0
  3. bundle.js +2 -0
  4. favicon.png +0 -0
  5. index.html +1 -1
  6. lib.js +0 -0
  7. preview.png +0 -0
analytics.js CHANGED
@@ -3,5 +3,5 @@
3
  m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
4
  })(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
5
 
6
- var ANALYTICS_ID = 'Add your own analytics ID here';
7
  ga('create', ANALYTICS_ID, 'auto');
 
3
  m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
4
  })(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
5
 
6
+ var ANALYTICS_ID = 'UA-46457317-4';
7
  ga('create', ANALYTICS_ID, 'auto');
bundle.css CHANGED
The diff for this file is too large to render. See raw diff
 
bundle.js CHANGED
@@ -0,0 +1,2 @@
 
 
 
1
+ !function e(t,n,r){function s(o,u){if(!n[o]){if(!t[o]){var a="function"==typeof require&&require;if(!u&&a)return a(o,!0);if(i)return i(o,!0);var f=new Error("Cannot find module '"+o+"'");throw f.code="MODULE_NOT_FOUND",f}var l=n[o]={exports:{}};t[o][0].call(l.exports,function(e){var n=t[o][1][e];return s(n?n:e)},l,l.exports,e,t,n,r)}return n[o].exports}for(var i="function"==typeof require&&require,o=0;o<r.length;o++)s(r[o]);s}({1:[function(require,module,exports){"use strict";function shuffle(array){for(var counter=array.length,temp=0,index=0;counter>0;)index=Math.floor(Math.random()*counter),counter--,temp=array[counter],array[counter]=array[index],array[index]=temp}function classifyTwoGaussData(numSamples,noise){function genGauss(cx,cy,label){for(var i=0;i<numSamples/2;i++){var x=normalRandom(cx,variance),y=normalRandom(cy,variance);points.push({x:x,y:y,label:label})}}var points=[],varianceScale=d3.scale.linear().domain([0,.5]).range([.5,4]),variance=varianceScale(noise);return genGauss(2,2,1),genGauss(-2,-2,-1),points}function regressPlane(numSamples,noise){for(var labelScale=d3.scale.linear().domain([-10,10]).range([-1,1]),getLabel=function(x,y){return labelScale(x+y)},points=[],i=0;i<numSamples;i++){var x=randUniform(-6,6),y=randUniform(-6,6),noiseX=randUniform(-6,6)*noise,noiseY=randUniform(-6,6)*noise,label=getLabel(x+noiseX,y+noiseY);points.push({x:x,y:y,label:label})}return points}function regressGaussian(numSamples,noise){function getLabel(x,y){var label=0;return gaussians.forEach(function(_a){var cx=_a[0],cy=_a[1],sign=_a[2],newLabel=sign*labelScale(dist({x:x,y:y},{x:cx,y:cy}));Math.abs(newLabel)>Math.abs(label)&&(label=newLabel)}),label}for(var points=[],labelScale=d3.scale.linear().domain([0,2]).range([1,0]).clamp(!0),gaussians=[[-4,2.5,1],[0,2.5,-1],[4,2.5,1],[-4,-2.5,-1],[0,-2.5,1],[4,-2.5,-1]],i=0;i<numSamples;i++){var x=randUniform(-6,6),y=randUniform(-6,6),noiseX=randUniform(-6,6)*noise,noiseY=randUniform(-6,6)*noise,label=getLabel(x+noiseX,y+noiseY);points.push({x:x,y:y,label:label})}return points}function classifySpiralData(numSamples,noise){function genSpiral(deltaT,label){for(var i=0;i<n;i++){var r=i/n*5,t=1.75*i/n*2*Math.PI+deltaT,x=r*Math.sin(t)+randUniform(-1,1)*noise,y=r*Math.cos(t)+randUniform(-1,1)*noise;points.push({x:x,y:y,label:label})}}var points=[],n=numSamples/2;return genSpiral(0,1),genSpiral(Math.PI,-1),points}function classifyCircleData(numSamples,noise){function getCircleLabel(p,center){return dist(p,center)<2.5?1:-1}for(var points=[],i=0;i<numSamples/2;i++){var r=randUniform(0,2.5),angle=randUniform(0,2*Math.PI),x=r*Math.sin(angle),y=r*Math.cos(angle),noiseX=randUniform(-5,5)*noise,noiseY=randUniform(-5,5)*noise,label=getCircleLabel({x:x+noiseX,y:y+noiseY},{x:0,y:0});points.push({x:x,y:y,label:label})}for(var i=0;i<numSamples/2;i++){var r=randUniform(3.5,5),angle=randUniform(0,2*Math.PI),x=r*Math.sin(angle),y=r*Math.cos(angle),noiseX=randUniform(-5,5)*noise,noiseY=randUniform(-5,5)*noise,label=getCircleLabel({x:x+noiseX,y:y+noiseY},{x:0,y:0});points.push({x:x,y:y,label:label})}return points}function classifyXORData(numSamples,noise){function getXORLabel(p){return p.x*p.y>=0?1:-1}for(var points=[],i=0;i<numSamples;i++){var x=randUniform(-5,5);x+=x>0?.3:-.3;var y=randUniform(-5,5);y+=y>0?.3:-.3;var noiseX=randUniform(-5,5)*noise,noiseY=randUniform(-5,5)*noise,label=getXORLabel({x:x+noiseX,y:y+noiseY});points.push({x:x,y:y,label:label})}return points}function randUniform(a,b){return Math.random()*(b-a)+a}function normalRandom(mean,variance){void 0===mean&&(mean=0),void 0===variance&&(variance=1);var v1,v2,s;do v1=2*Math.random()-1,v2=2*Math.random()-1,s=v1*v1+v2*v2;while(s>1);var result=Math.sqrt(-2*Math.log(s)/s)*v1;return mean+Math.sqrt(variance)*result}function dist(a,b){var dx=a.x-b.x,dy=a.y-b.y;return Math.sqrt(dx*dx+dy*dy)}exports.shuffle=shuffle,exports.classifyTwoGaussData=classifyTwoGaussData,exports.regressPlane=regressPlane,exports.regressGaussian=regressGaussian,exports.classifySpiralData=classifySpiralData,exports.classifyCircleData=classifyCircleData,exports.classifyXORData=classifyXORData},{}],2:[function(require,module,exports){"use strict";function reduceMatrix(matrix,factor){if(matrix.length!==matrix[0].length)throw new Error("The provided matrix must be a square matrix");if(matrix.length%factor!=0)throw new Error("The width/height of the matrix must be divisible by the reduction factor");for(var result=new Array(matrix.length/factor),i=0;i<matrix.length;i+=factor){result[i/factor]=new Array(matrix.length/factor);for(var j=0;j<matrix.length;j+=factor){for(var avg=0,k=0;k<factor;k++)for(var l=0;l<factor;l++)avg+=matrix[i+k][j+l];avg/=factor*factor,result[i/factor][j/factor]=avg}}return result}var HeatMap=function(){function HeatMap(width,numSamples,xDomain,yDomain,container,userSettings){this.settings={showAxes:!1,noSvg:!1},this.numSamples=numSamples;var height=width,padding=userSettings.showAxes?20:0;if(null!=userSettings)for(var prop in userSettings)this.settings[prop]=userSettings[prop];this.xScale=d3.scale.linear().domain(xDomain).range([0,width-2*padding]),this.yScale=d3.scale.linear().domain(yDomain).range([height-2*padding,0]);var tmpScale=d3.scale.linear().domain([0,.5,1]).range(["#f59322","#e8eaeb","#0877bd"]).clamp(!0),colors=d3.range(0,1+1e-9,1/30).map(function(a){return tmpScale(a)});if(this.color=d3.scale.quantize().domain([-1,1]).range(colors),container=container.append("div").style({width:width+"px",height:height+"px",position:"relative",top:"-"+padding+"px",left:"-"+padding+"px"}),this.canvas=container.append("canvas").attr("width",numSamples).attr("height",numSamples).style("width",width-2*padding+"px").style("height",height-2*padding+"px").style("position","absolute").style("top",padding+"px").style("left",padding+"px"),this.settings.noSvg||(this.svg=container.append("svg").attr({width:width,height:height}).style({position:"absolute",left:"0",top:"0"}).append("g").attr("transform","translate("+padding+","+padding+")"),this.svg.append("g").attr("class","train"),this.svg.append("g").attr("class","test")),this.settings.showAxes){var xAxis=d3.svg.axis().scale(this.xScale).orient("bottom"),yAxis=d3.svg.axis().scale(this.yScale).orient("right");this.svg.append("g").attr("class","x axis").attr("transform","translate(0,"+(height-2*padding)+")").call(xAxis),this.svg.append("g").attr("class","y axis").attr("transform","translate("+(width-2*padding)+",0)").call(yAxis)}}return HeatMap.prototype.updateTestPoints=function(points){if(this.settings.noSvg)throw Error("Can't add points since noSvg=true");this.updateCircles(this.svg.select("g.test"),points)},HeatMap.prototype.updatePoints=function(points){if(this.settings.noSvg)throw Error("Can't add points since noSvg=true");this.updateCircles(this.svg.select("g.train"),points)},HeatMap.prototype.updateBackground=function(data,discretize){var dx=data[0].length,dy=data.length;if(dx!==this.numSamples||dy!==this.numSamples)throw new Error("The provided data matrix must be of size numSamples X numSamples");for(var context=this.canvas.node().getContext("2d"),image=context.createImageData(dx,dy),y=0,p=-1;y<dy;++y)for(var x=0;x<dx;++x){var value=data[x][y];discretize&&(value=value>=0?1:-1);var c=d3.rgb(this.color(value));image.data[++p]=c.r,image.data[++p]=c.g,image.data[++p]=c.b,image.data[++p]=160}context.putImageData(image,0,0)},HeatMap.prototype.updateCircles=function(container,points){var _this=this,xDomain=this.xScale.domain(),yDomain=this.yScale.domain();points=points.filter(function(p){return p.x>=xDomain[0]&&p.x<=xDomain[1]&&p.y>=yDomain[0]&&p.y<=yDomain[1]});var selection=container.selectAll("circle").data(points);selection.enter().append("circle").attr("r",3),selection.attr({cx:function(d){return _this.xScale(d.x)},cy:function(d){return _this.yScale(d.y)}}).style("fill",function(d){return _this.color(d.label)}),selection.exit().remove()},HeatMap}();exports.HeatMap=HeatMap,exports.reduceMatrix=reduceMatrix},{}],3:[function(require,module,exports){"use strict";var AppendingLineChart=function(){function AppendingLineChart(container,lineColors){this.data=[],this.minY=Number.MAX_VALUE,this.maxY=Number.MIN_VALUE,this.lineColors=lineColors,this.numLines=lineColors.length;var node=container.node(),totalWidth=node.offsetWidth,totalHeight=node.offsetHeight,margin={top:2,right:0,bottom:2,left:2},width=totalWidth-margin.left-margin.right,height=totalHeight-margin.top-margin.bottom;this.xScale=d3.scale.linear().domain([0,0]).range([0,width]),this.yScale=d3.scale.linear().domain([0,0]).range([height,0]),this.svg=container.append("svg").attr("width",width+margin.left+margin.right).attr("height",height+margin.top+margin.bottom).append("g").attr("transform","translate("+margin.left+","+margin.top+")"),this.paths=new Array(this.numLines);for(var i=0;i<this.numLines;i++)this.paths[i]=this.svg.append("path").attr("class","line").style({fill:"none",stroke:lineColors[i],"stroke-width":"1.5px"})}return AppendingLineChart.prototype.reset=function(){this.data=[],this.redraw(),this.minY=Number.MAX_VALUE,this.maxY=Number.MIN_VALUE},AppendingLineChart.prototype.addDataPoint=function(dataPoint){var _this=this;if(dataPoint.length!==this.numLines)throw Error("Length of dataPoint must equal number of lines");dataPoint.forEach(function(y){_this.minY=Math.min(_this.minY,y),_this.maxY=Math.max(_this.maxY,y)}),this.data.push({x:this.data.length+1,y:dataPoint}),this.redraw()},AppendingLineChart.prototype.redraw=function(){var _this=this;this.xScale.domain([1,this.data.length]),this.yScale.domain([this.minY,this.maxY]);for(var getPathMap=function(lineIndex){return d3.svg.line().x(function(d){return _this.xScale(d.x)}).y(function(d){return _this.yScale(d.y[lineIndex])})},i=0;i<this.numLines;i++)this.paths[i].datum(this.data).attr("d",getPathMap(i))},AppendingLineChart}();exports.AppendingLineChart=AppendingLineChart},{}],4:[function(require,module,exports){"use strict";function buildNetwork(networkShape,activation,outputActivation,regularization,inputIds,initZero){for(var numLayers=networkShape.length,id=1,network=[],layerIdx=0;layerIdx<numLayers;layerIdx++){var isOutputLayer=layerIdx===numLayers-1,isInputLayer=0===layerIdx,currentLayer=[];network.push(currentLayer);for(var numNodes=networkShape[layerIdx],i=0;i<numNodes;i++){var nodeId=id.toString();isInputLayer?nodeId=inputIds[i]:id++;var node=new Node(nodeId,isOutputLayer?outputActivation:activation,initZero);if(currentLayer.push(node),layerIdx>=1)for(var j=0;j<network[layerIdx-1].length;j++){var prevNode=network[layerIdx-1][j],link=new Link(prevNode,node,regularization,initZero);prevNode.outputs.push(link),node.inputLinks.push(link)}}}return network}function forwardProp(network,inputs){var inputLayer=network[0];if(inputs.length!==inputLayer.length)throw new Error("The number of inputs must match the number of nodes in the input layer");for(var i=0;i<inputLayer.length;i++){var node=inputLayer[i];node.output=inputs[i]}for(var layerIdx=1;layerIdx<network.length;layerIdx++)for(var currentLayer=network[layerIdx],i=0;i<currentLayer.length;i++){var node=currentLayer[i];node.updateOutput()}return network[network.length-1][0].output}function backProp(network,target,errorFunc){var outputNode=network[network.length-1][0];outputNode.outputDer=errorFunc.der(outputNode.output,target);for(var layerIdx=network.length-1;layerIdx>=1;layerIdx--){for(var currentLayer=network[layerIdx],i=0;i<currentLayer.length;i++){var node=currentLayer[i];node.inputDer=node.outputDer*node.activation.der(node.totalInput),node.accInputDer+=node.inputDer,node.numAccumulatedDers++}for(var i=0;i<currentLayer.length;i++)for(var node=currentLayer[i],j=0;j<node.inputLinks.length;j++){var link=node.inputLinks[j];link.isDead||(link.errorDer=node.inputDer*link.source.output,link.accErrorDer+=link.errorDer,link.numAccumulatedDers++)}if(1!==layerIdx)for(var prevLayer=network[layerIdx-1],i=0;i<prevLayer.length;i++){var node=prevLayer[i];node.outputDer=0;for(var j=0;j<node.outputs.length;j++){var output=node.outputs[j];node.outputDer+=output.weight*output.dest.inputDer}}}}function updateWeights(network,learningRate,regularizationRate){for(var layerIdx=1;layerIdx<network.length;layerIdx++)for(var currentLayer=network[layerIdx],i=0;i<currentLayer.length;i++){var node=currentLayer[i];node.numAccumulatedDers>0&&(node.bias-=learningRate*node.accInputDer/node.numAccumulatedDers,node.accInputDer=0,node.numAccumulatedDers=0);for(var j=0;j<node.inputLinks.length;j++){var link=node.inputLinks[j];if(!link.isDead){var regulDer=link.regularization?link.regularization.der(link.weight):0;if(link.numAccumulatedDers>0){link.weight=link.weight-learningRate/link.numAccumulatedDers*link.accErrorDer;var newLinkWeight=link.weight-learningRate*regularizationRate*regulDer;link.regularization===RegularizationFunction.L1&&link.weight*newLinkWeight<0?(link.weight=0,link.isDead=!0):link.weight=newLinkWeight,link.accErrorDer=0,link.numAccumulatedDers=0}}}}}function forEachNode(network,ignoreInputs,accessor){for(var layerIdx=ignoreInputs?1:0;layerIdx<network.length;layerIdx++)for(var currentLayer=network[layerIdx],i=0;i<currentLayer.length;i++){var node=currentLayer[i];accessor(node)}}function getOutputNode(network){return network[network.length-1][0]}var Node=function(){function Node(id,activation,initZero){this.inputLinks=[],this.bias=.1,this.outputs=[],this.outputDer=0,this.inputDer=0,this.accInputDer=0,this.numAccumulatedDers=0,this.id=id,this.activation=activation,initZero&&(this.bias=0)}return Node.prototype.updateOutput=function(){this.totalInput=this.bias;for(var j=0;j<this.inputLinks.length;j++){var link=this.inputLinks[j];this.totalInput+=link.weight*link.source.output}return this.output=this.activation.output(this.totalInput),this.output},Node}();exports.Node=Node;var Errors=function(){function Errors(){}return Errors}();Errors.SQUARE={error:function(output,target){return.5*Math.pow(output-target,2)},der:function(output,target){return output-target}},exports.Errors=Errors,Math.tanh=Math.tanh||function(x){if(x===1/0)return 1;if(x===-(1/0))return-1;var e2x=Math.exp(2*x);return(e2x-1)/(e2x+1)};var Activations=function(){function Activations(){}return Activations}();Activations.TANH={output:function(x){return Math.tanh(x)},der:function(x){var output=Activations.TANH.output(x);return 1-output*output}},Activations.RELU={output:function(x){return Math.max(0,x)},der:function(x){return x<=0?0:1}},Activations.SIGMOID={output:function(x){return 1/(1+Math.exp(-x))},der:function(x){var output=Activations.SIGMOID.output(x);return output*(1-output)}},Activations.LINEAR={output:function(x){return x},der:function(x){return 1}},exports.Activations=Activations;var RegularizationFunction=function(){function RegularizationFunction(){}return RegularizationFunction}();RegularizationFunction.L1={output:function(w){return Math.abs(w)},der:function(w){return w<0?-1:w>0?1:0}},RegularizationFunction.L2={output:function(w){return.5*w*w},der:function(w){return w}},exports.RegularizationFunction=RegularizationFunction;var Link=function(){function Link(source,dest,regularization,initZero){this.weight=Math.random()-.5,this.isDead=!1,this.errorDer=0,this.accErrorDer=0,this.numAccumulatedDers=0,this.id=source.id+"-"+dest.id,this.source=source,this.dest=dest,this.regularization=regularization,initZero&&(this.weight=0)}return Link}();exports.Link=Link,exports.buildNetwork=buildNetwork,exports.forwardProp=forwardProp,exports.backProp=backProp,exports.updateWeights=updateWeights,exports.forEachNode=forEachNode,exports.getOutputNode=getOutputNode},{}],5:[function(require,module,exports){"use strict";function scrollTween(offset){return function(){var i=d3.interpolateNumber(window.pageYOffset||document.documentElement.scrollTop,offset);return function(t){scrollTo(0,i(t))}}}function makeGUI(){d3.select("#reset-button").on("click",function(){reset(),userHasInteracted(),d3.select("#play-pause-button")}),d3.select("#play-pause-button").on("click",function(){userHasInteracted(),player.playOrPause()}),player.onPlayPause(function(isPlaying){d3.select("#play-pause-button").classed("playing",isPlaying)}),d3.select("#next-step-button").on("click",function(){player.pause(),userHasInteracted(),0===iter&&simulationStarted(),oneStep()}),d3.select("#data-regen-button").on("click",function(){generateData(),parametersChanged=!0});var dataThumbnails=d3.selectAll("canvas[data-dataset]");dataThumbnails.on("click",function(){var newDataset=state_1.datasets[this.dataset.dataset];newDataset!==state.dataset&&(state.dataset=newDataset,dataThumbnails.classed("selected",!1),d3.select(this).classed("selected",!0),generateData(),parametersChanged=!0,reset())});var datasetKey=state_1.getKeyFromValue(state_1.datasets,state.dataset);d3.select("canvas[data-dataset="+datasetKey+"]").classed("selected",!0);var regDataThumbnails=d3.selectAll("canvas[data-regDataset]");regDataThumbnails.on("click",function(){var newDataset=state_1.regDatasets[this.dataset.regdataset];newDataset!==state.regDataset&&(state.regDataset=newDataset,regDataThumbnails.classed("selected",!1),d3.select(this).classed("selected",!0),generateData(),parametersChanged=!0,reset())});var regDatasetKey=state_1.getKeyFromValue(state_1.regDatasets,state.regDataset);d3.select("canvas[data-regDataset="+regDatasetKey+"]").classed("selected",!0),d3.select("#add-layers").on("click",function(){state.numHiddenLayers>=6||(state.networkShape[state.numHiddenLayers]=2,state.numHiddenLayers++,parametersChanged=!0,reset())}),d3.select("#remove-layers").on("click",function(){state.numHiddenLayers<=0||(state.numHiddenLayers--,state.networkShape.splice(state.numHiddenLayers),parametersChanged=!0,reset())}),d3.select("#show-test-data").on("change",function(){state.showTestData=this.checked,state.serialize(),userHasInteracted(),heatMap.updateTestPoints(state.showTestData?testData:[])}).property("checked",state.showTestData),d3.select("#discretize").on("change",function(){state.discretize=this.checked,state.serialize(),userHasInteracted(),updateUI()}).property("checked",state.discretize),d3.select("#percTrainData").on("input",function(){state.percTrainData=this.value,d3.select("label[for='percTrainData'] .value").text(this.value),generateData(),parametersChanged=!0,reset()}).property("value",state.percTrainData),d3.select("label[for='percTrainData'] .value").text(state.percTrainData),d3.select("#noise").on("input",function(){state.noise=this.value,d3.select("label[for='noise'] .value").text(this.value),generateData(),parametersChanged=!0,reset()}).property("value",state.noise),d3.select("label[for='noise'] .value").text(state.noise),d3.select("#batchSize").on("input",function(){state.batchSize=this.value,d3.select("label[for='batchSize'] .value").text(this.value),parametersChanged=!0,reset()}).property("value",state.batchSize),d3.select("label[for='batchSize'] .value").text(state.batchSize),d3.select("#activations").on("change",function(){state.activation=state_1.activations[this.value],parametersChanged=!0,reset()}).property("value",state_1.getKeyFromValue(state_1.activations,state.activation)),d3.select("#learningRate").on("change",function(){state.learningRate=+this.value,state.serialize(),userHasInteracted(),parametersChanged=!0}).property("value",state.learningRate),d3.select("#regularizations").on("change",function(){state.regularization=state_1.regularizations[this.value],parametersChanged=!0,reset()}).property("value",state_1.getKeyFromValue(state_1.regularizations,state.regularization)),d3.select("#regularRate").on("change",function(){state.regularizationRate=+this.value,parametersChanged=!0,reset()}).property("value",state.regularizationRate),d3.select("#problem").on("change",function(){state.problem=state_1.problems[this.value],generateData(),drawDatasetThumbnails(),parametersChanged=!0,reset()}).property("value",state_1.getKeyFromValue(state_1.problems,state.problem));var x=d3.scale.linear().domain([-1,1]).range([0,144]),xAxis=d3.svg.axis().scale(x).orient("bottom").tickValues([-1,0,1]).tickFormat(d3.format("d"));d3.select("#colormap g.core").append("g").attr("class","x axis").attr("transform","translate(0,10)").call(xAxis),window.addEventListener("resize",function(){var newWidth=document.querySelector("#main-part").getBoundingClientRect().width;newWidth!==mainWidth&&(mainWidth=newWidth,drawNetwork(network),updateUI(!0))}),state.hideText&&(d3.select("#article-text").style("display","none"),d3.select("div.more").style("display","none"),d3.select("header").style("display","none"))}function updateBiasesUI(network){nn.forEachNode(network,!0,function(node){d3.select("rect#bias-"+node.id).style("fill",colorScale(node.bias))})}function updateWeightsUI(network,container){for(var layerIdx=1;layerIdx<network.length;layerIdx++)for(var currentLayer=network[layerIdx],i=0;i<currentLayer.length;i++)for(var node=currentLayer[i],j=0;j<node.inputLinks.length;j++){var link=node.inputLinks[j];container.select("#link"+link.source.id+"-"+link.dest.id).style({"stroke-dashoffset":-iter/3,"stroke-width":linkWidthScale(Math.abs(link.weight)),stroke:colorScale(link.weight)}).datum(link)}}function drawNode(cx,cy,nodeId,isInput,container,node){var x=cx-15,y=cy-15,nodeGroup=container.append("g").attr({class:"node",id:"node"+nodeId,transform:"translate("+x+","+y+")"});nodeGroup.append("rect").attr({x:0,y:0,width:30,height:30});var activeOrNotClass=state[nodeId]?"active":"inactive";if(isInput){var label=null!=INPUTS[nodeId].label?INPUTS[nodeId].label:nodeId,text=nodeGroup.append("text").attr({class:"main-label",x:-10,y:15,"text-anchor":"end"});if(/[_^]/.test(label)){for(var myRe=/(.*?)([_^])(.)/g,myArray=void 0,lastIndex=void 0;null!=(myArray=myRe.exec(label));){lastIndex=myRe.lastIndex;var prefix=myArray[1],sep=myArray[2],suffix=myArray[3];prefix&&text.append("tspan").text(prefix),text.append("tspan").attr("baseline-shift","_"===sep?"sub":"super").style("font-size","9px").text(suffix)}label.substring(lastIndex)&&text.append("tspan").text(label.substring(lastIndex))}else text.append("tspan").text(label);nodeGroup.classed(activeOrNotClass,!0)}isInput||nodeGroup.append("rect").attr({id:"bias-"+nodeId,x:-7,y:28,width:5,height:5}).on("mouseenter",function(){updateHoverCard(HoverType.BIAS,node,d3.mouse(container.node()))}).on("mouseleave",function(){updateHoverCard(null)});var div=d3.select("#network").insert("div",":first-child").attr({id:"canvas-"+nodeId,class:"canvas"}).style({position:"absolute",left:x+3+"px",top:y+3+"px"}).on("mouseenter",function(){selectedNodeId=nodeId,div.classed("hovered",!0),nodeGroup.classed("hovered",!0),updateDecisionBoundary(network,!1),heatMap.updateBackground(boundary[nodeId],state.discretize)}).on("mouseleave",function(){selectedNodeId=null,div.classed("hovered",!1),nodeGroup.classed("hovered",!1),updateDecisionBoundary(network,!1),heatMap.updateBackground(boundary[nn.getOutputNode(network).id],state.discretize)});isInput&&(div.on("click",function(){state[nodeId]=!state[nodeId],parametersChanged=!0,reset()}),div.style("cursor","pointer")),isInput&&div.classed(activeOrNotClass,!0);var nodeHeatMap=new heatmap_1.HeatMap(30,10,xDomain,xDomain,div,{noSvg:!0});div.datum({heatmap:nodeHeatMap,id:nodeId})}function drawNetwork(network){var svg=d3.select("#svg");svg.select("g.core").remove(),d3.select("#network").selectAll("div.canvas").remove(),d3.select("#network").selectAll("div.plus-minus-neurons").remove();var co=d3.select(".column.output").node(),cf=d3.select(".column.features").node(),width=co.offsetLeft-cf.offsetLeft;svg.attr("width",width);var node2coord={},container=svg.append("g").classed("core",!0).attr("transform","translate(3,3)"),numLayers=network.length,layerScale=d3.scale.ordinal().domain(d3.range(1,numLayers-1)).rangePoints([118,width-30],.7),nodeIndexScale=function(nodeIndex){return 55*nodeIndex},calloutThumb=d3.select(".callout.thumbnail").style("display","none"),calloutWeights=d3.select(".callout.weights").style("display","none"),idWithCallout=null,targetIdWithCallout=null,cx=65,nodeIds=Object.keys(INPUTS),maxY=nodeIndexScale(nodeIds.length);nodeIds.forEach(function(nodeId,i){var cy=nodeIndexScale(i)+15;node2coord[nodeId]={cx:cx,cy:cy},drawNode(cx,cy,nodeId,!0,container)});for(var layerIdx=1;layerIdx<numLayers-1;layerIdx++){var numNodes=network[layerIdx].length,cx_1=layerScale(layerIdx)+15;maxY=Math.max(maxY,nodeIndexScale(numNodes)),addPlusMinusControl(layerScale(layerIdx),layerIdx);for(var i=0;i<numNodes;i++){var node_1=network[layerIdx][i],cy_1=nodeIndexScale(i)+15;node2coord[node_1.id]={cx:cx_1,cy:cy_1},drawNode(cx_1,cy_1,node_1.id,!1,container,node_1);var numNodes_1=network[layerIdx].length,nextNumNodes=network[layerIdx+1].length;null==idWithCallout&&i===numNodes_1-1&&nextNumNodes<=numNodes_1&&(calloutThumb.style({display:null,top:23+cy_1+"px",left:cx_1+"px"}),idWithCallout=node_1.id);for(var j=0;j<node_1.inputLinks.length;j++){var link=node_1.inputLinks[j],path=drawLink(link,node2coord,network,container,0===j,j,node_1.inputLinks.length).node(),prevLayer=network[layerIdx-1],lastNodePrevLayer=prevLayer[prevLayer.length-1];if(null==targetIdWithCallout&&i===numNodes_1-1&&link.source.id===lastNodePrevLayer.id&&(link.source.id!==idWithCallout||numLayers<=5)&&link.dest.id!==idWithCallout&&prevLayer.length>=numNodes_1){var midPoint=path.getPointAtLength(.7*path.getTotalLength());calloutWeights.style({display:null,top:midPoint.y+5+"px",left:midPoint.x+3+"px"}),targetIdWithCallout=link.dest.id}}}}cx=width+15;var node=network[numLayers-1][0],cy=nodeIndexScale(0)+15;node2coord[node.id]={cx:cx,cy:cy};for(var i=0;i<node.inputLinks.length;i++){var link=node.inputLinks[i];drawLink(link,node2coord,network,container,0===i,i,node.inputLinks.length)}svg.attr("height",maxY);var height=Math.max(getRelativeHeight(calloutThumb),getRelativeHeight(calloutWeights),getRelativeHeight(d3.select("#network")));d3.select(".column.features").style("height",height+"px")}function getRelativeHeight(selection){var node=selection.node();return node.offsetHeight+node.offsetTop}function addPlusMinusControl(x,layerIdx){var div=d3.select("#network").append("div").classed("plus-minus-neurons",!0).style("left",x-10+"px"),i=layerIdx-1,firstRow=div.append("div").attr("class","ui-numNodes"+layerIdx);firstRow.append("button").attr("class","mdl-button mdl-js-button mdl-button--icon").on("click",function(){state.networkShape[i]>=8||(state.networkShape[i]++,parametersChanged=!0,reset())}).append("i").attr("class","material-icons").text("add"),firstRow.append("button").attr("class","mdl-button mdl-js-button mdl-button--icon").on("click",function(){state.networkShape[i]<=1||(state.networkShape[i]--,parametersChanged=!0,reset())}).append("i").attr("class","material-icons").text("remove");var suffix=state.networkShape[i]>1?"s":"";div.append("div").text(state.networkShape[i]+" neuron"+suffix)}function updateHoverCard(type,nodeOrLink,coordinates){var hovercard=d3.select("#hovercard");if(null==type)return hovercard.style("display","none"),void d3.select("#svg").on("click",null);d3.select("#svg").on("click",function(){hovercard.select(".value").style("display","none");var input=hovercard.select("input");input.style("display",null),input.on("input",function(){null!=this.value&&""!==this.value&&(type===HoverType.WEIGHT?nodeOrLink.weight=+this.value:nodeOrLink.bias=+this.value,updateUI())}),input.on("keypress",function(){13===d3.event.keyCode&&updateHoverCard(type,nodeOrLink,coordinates)}),input.node().focus()});var value=type===HoverType.WEIGHT?nodeOrLink.weight:nodeOrLink.bias,name=type===HoverType.WEIGHT?"Weight":"Bias";hovercard.style({left:coordinates[0]+20+"px",top:coordinates[1]+"px",display:"block"}),hovercard.select(".type").text(name),hovercard.select(".value").style("display",null).text(value.toPrecision(2)),hovercard.select("input").property("value",value.toPrecision(2)).style("display","none")}function drawLink(input,node2coord,network,container,isFirst,index,length){var line=container.insert("path",":first-child"),source=node2coord[input.source.id],dest=node2coord[input.dest.id],datum={source:{y:source.cx+15+2,x:source.cy},target:{y:dest.cx-15,x:dest.cy+(index-(length-1)/2)/length*12}},diagonal=d3.svg.diagonal().projection(function(d){return[d.y,d.x]});return line.attr({"marker-start":"url(#markerArrow)",class:"link",id:"link"+input.source.id+"-"+input.dest.id,d:diagonal(datum,0)}),container.append("path").attr("d",diagonal(datum,0)).attr("class","link-hover").on("mouseenter",function(){updateHoverCard(HoverType.WEIGHT,input,d3.mouse(this))}).on("mouseleave",function(){updateHoverCard(null)}),line}function updateDecisionBoundary(network,firstTime){if(firstTime){boundary={},nn.forEachNode(network,!0,function(node){boundary[node.id]=new Array(100)});for(var nodeId in INPUTS)boundary[nodeId]=new Array(100)}var xScale=d3.scale.linear().domain([0,99]).range(xDomain),yScale=d3.scale.linear().domain([99,0]).range(xDomain),i=0,j=0;for(i=0;i<100;i++){if(firstTime){nn.forEachNode(network,!0,function(node){boundary[node.id][i]=new Array(100)});for(var nodeId in INPUTS)boundary[nodeId][i]=new Array(100)}for(j=0;j<100;j++){var x=xScale(i),y=yScale(j),input=constructInput(x,y);if(nn.forwardProp(network,input),nn.forEachNode(network,!0,function(node){boundary[node.id][i][j]=node.output}),firstTime)for(var nodeId in INPUTS)boundary[nodeId][i][j]=INPUTS[nodeId].f(x,y)}}}function getLoss(network,dataPoints){for(var loss=0,i=0;i<dataPoints.length;i++){var dataPoint=dataPoints[i],input=constructInput(dataPoint.x,dataPoint.y),output=nn.forwardProp(network,input);loss+=nn.Errors.SQUARE.error(output,dataPoint.label)}return loss/dataPoints.length}function updateUI(firstStep){function zeroPad(n){return("000000"+n).slice(-"000000".length)}function addCommas(s){return s.replace(/\B(?=(\d{3})+(?!\d))/g,",")}function humanReadable(n){return n.toFixed(3)}void 0===firstStep&&(firstStep=!1),updateWeightsUI(network,d3.select("g.core")),updateBiasesUI(network),updateDecisionBoundary(network,firstStep);var selectedId=null!=selectedNodeId?selectedNodeId:nn.getOutputNode(network).id;heatMap.updateBackground(boundary[selectedId],state.discretize),d3.select("#network").selectAll("div.canvas").each(function(data){data.heatmap.updateBackground(heatmap_1.reduceMatrix(boundary[data.id],10),state.discretize)}),d3.select("#loss-train").text(humanReadable(lossTrain)),d3.select("#loss-test").text(humanReadable(lossTest)),d3.select("#iter-number").text(addCommas(zeroPad(iter))),lineChart.addDataPoint([lossTrain,lossTest])}function constructInputIds(){var result=[];for(var inputName in INPUTS)state[inputName]&&result.push(inputName);return result}function constructInput(x,y){var input=[];for(var inputName in INPUTS)state[inputName]&&input.push(INPUTS[inputName].f(x,y));return input}function oneStep(){iter++,trainData.forEach(function(point,i){var input=constructInput(point.x,point.y);nn.forwardProp(network,input),nn.backProp(network,point.label,nn.Errors.SQUARE),(i+1)%state.batchSize==0&&nn.updateWeights(network,state.learningRate,state.regularizationRate)}),lossTrain=getLoss(network,trainData),lossTest=getLoss(network,testData),updateUI()}function getOutputWeights(network){for(var weights=[],layerIdx=0;layerIdx<network.length-1;layerIdx++)for(var currentLayer=network[layerIdx],i=0;i<currentLayer.length;i++)for(var node=currentLayer[i],j=0;j<node.outputs.length;j++){var output=node.outputs[j];weights.push(output.weight)}return weights}function reset(onStartup){void 0===onStartup&&(onStartup=!1),lineChart.reset(),state.serialize(),onStartup||userHasInteracted(),player.pause();var suffix=1!==state.numHiddenLayers?"s":"";d3.select("#layers-label").text("Hidden layer"+suffix),d3.select("#num-layers").text(state.numHiddenLayers),iter=0;var numInputs=constructInput(0,0).length,shape=[numInputs].concat(state.networkShape).concat([1]),outputActivation=state.problem===state_1.Problem.REGRESSION?nn.Activations.LINEAR:nn.Activations.TANH
2
+ ;network=nn.buildNetwork(shape,state.activation,outputActivation,state.regularization,constructInputIds(),state.initZero),lossTrain=getLoss(network,trainData),lossTest=getLoss(network,testData),drawNetwork(network),updateUI(!0)}function initTutorial(){if(null!=state.tutorial&&""!==state.tutorial&&!state.hideText){d3.selectAll("article div.l--body").remove();var tutorial=d3.select("article").append("div").attr("class","l--body");d3.html("tutorials/"+state.tutorial+".html",function(err,htmlFragment){if(err)throw err;tutorial.node().appendChild(htmlFragment);var title=tutorial.select("title");title.size()&&(d3.select("header h1").style({"margin-top":"20px","margin-bottom":"20px"}).text(title.text()),document.title=title.text())})}}function drawDatasetThumbnails(){function renderThumbnail(canvas,dataGenerator){canvas.setAttribute("width",100),canvas.setAttribute("height",100);var context=canvas.getContext("2d");dataGenerator(200,0).forEach(function(d){context.fillStyle=colorScale(d.label),context.fillRect(100*(d.x+6)/12,100*(d.y+6)/12,4,4)}),d3.select(canvas.parentNode).style("display",null)}if(d3.selectAll(".dataset").style("display","none"),state.problem===state_1.Problem.CLASSIFICATION)for(var dataset in state_1.datasets){var canvas=document.querySelector("canvas[data-dataset="+dataset+"]"),dataGenerator=state_1.datasets[dataset];renderThumbnail(canvas,dataGenerator)}if(state.problem===state_1.Problem.REGRESSION)for(var regDataset in state_1.regDatasets){var canvas=document.querySelector("canvas[data-regDataset="+regDataset+"]"),dataGenerator=state_1.regDatasets[regDataset];renderThumbnail(canvas,dataGenerator)}}function hideControls(){var hiddenProps=state.getHiddenProps();hiddenProps.forEach(function(prop){var controls=d3.selectAll(".ui-"+prop);0===controls.size()&&console.warn("0 html elements found with class .ui-"+prop),controls.style("display","none")});var hideControls=d3.select(".hide-controls");HIDABLE_CONTROLS.forEach(function(_a){var text=_a[0],id=_a[1],label=hideControls.append("label").attr("class","mdl-checkbox mdl-js-checkbox mdl-js-ripple-effect"),input=label.append("input").attr({type:"checkbox",class:"mdl-checkbox__input"});hiddenProps.indexOf(id)===-1&&input.attr("checked","true"),input.on("change",function(){state.setHideProperty(id,!this.checked),state.serialize(),userHasInteracted(),d3.select(".hide-controls-link").attr("href",window.location.href)}),label.append("span").attr("class","mdl-checkbox__label label").text(text)}),d3.select(".hide-controls-link").attr("href",window.location.href)}function generateData(firstTime){void 0===firstTime&&(firstTime=!1),firstTime||(state.seed=Math.random().toFixed(5),state.serialize(),userHasInteracted()),Math.seedrandom(state.seed);var numSamples=state.problem===state_1.Problem.REGRESSION?1200:500,generator=state.problem===state_1.Problem.CLASSIFICATION?state.dataset:state.regDataset,data=generator(numSamples,state.noise/100);dataset_1.shuffle(data);var splitIndex=Math.floor(data.length*state.percTrainData/100);trainData=data.slice(0,splitIndex),testData=data.slice(splitIndex),heatMap.updatePoints(trainData),heatMap.updateTestPoints(state.showTestData?testData:[])}function userHasInteracted(){if(firstInteraction){firstInteraction=!1;var page="index";null!=state.tutorial&&""!==state.tutorial&&(page="/v/tutorials/"+state.tutorial),ga("set","page",page),ga("send","pageview",{sessionControl:"start"})}}function simulationStarted(){ga("send",{hitType:"event",eventCategory:"Starting Simulation",eventAction:parametersChanged?"changed":"unchanged",eventLabel:null==state.tutorial?"":state.tutorial}),parametersChanged=!1}var mainWidth,nn=require("./nn"),heatmap_1=require("./heatmap"),state_1=require("./state"),dataset_1=require("./dataset"),linechart_1=require("./linechart");d3.select(".more button").on("click",function(){d3.transition().duration(1e3).tween("scroll",scrollTween(800))});var HoverType;!function(HoverType){HoverType[HoverType.BIAS=0]="BIAS",HoverType[HoverType.WEIGHT=1]="WEIGHT"}(HoverType||(HoverType={}));var INPUTS={x:{f:function(x,y){return x},label:"X_1"},y:{f:function(x,y){return y},label:"X_2"},xSquared:{f:function(x,y){return x*x},label:"X_1^2"},ySquared:{f:function(x,y){return y*y},label:"X_2^2"},xTimesY:{f:function(x,y){return x*y},label:"X_1X_2"},sinX:{f:function(x,y){return Math.sin(x)},label:"sin(X_1)"},sinY:{f:function(x,y){return Math.sin(y)},label:"sin(X_2)"}},HIDABLE_CONTROLS=[["Show test data","showTestData"],["Discretize output","discretize"],["Play button","playButton"],["Step button","stepButton"],["Reset button","resetButton"],["Learning rate","learningRate"],["Activation","activation"],["Regularization","regularization"],["Regularization rate","regularizationRate"],["Problem type","problem"],["Which dataset","dataset"],["Ratio train data","percTrainData"],["Noise level","noise"],["Batch size","batchSize"],["# of hidden layers","numHiddenLayers"]],Player=function(){function Player(){this.timerIndex=0,this.isPlaying=!1,this.callback=null}return Player.prototype.playOrPause=function(){this.isPlaying?(this.isPlaying=!1,this.pause()):(this.isPlaying=!0,0===iter&&simulationStarted(),this.play())},Player.prototype.onPlayPause=function(callback){this.callback=callback},Player.prototype.play=function(){this.pause(),this.isPlaying=!0,this.callback&&this.callback(this.isPlaying),this.start(this.timerIndex)},Player.prototype.pause=function(){this.timerIndex++,this.isPlaying=!1,this.callback&&this.callback(this.isPlaying)},Player.prototype.start=function(localTimerIndex){var _this=this;d3.timer(function(){return localTimerIndex<_this.timerIndex||(oneStep(),!1)},0)},Player}(),state=state_1.State.deserializeState();state.getHiddenProps().forEach(function(prop){prop in INPUTS&&delete INPUTS[prop]});var boundary={},selectedNodeId=null,xDomain=[-6,6],heatMap=new heatmap_1.HeatMap(300,100,xDomain,xDomain,d3.select("#heatmap"),{showAxes:!0}),linkWidthScale=d3.scale.linear().domain([0,5]).range([1,10]).clamp(!0),colorScale=d3.scale.linear().domain([-1,0,1]).range(["#f59322","#e8eaeb","#0877bd"]).clamp(!0),iter=0,trainData=[],testData=[],network=null,lossTrain=0,lossTest=0,player=new Player,lineChart=new linechart_1.AppendingLineChart(d3.select("#linechart"),["#777","black"]);exports.getOutputWeights=getOutputWeights;var firstInteraction=!0,parametersChanged=!1;drawDatasetThumbnails(),initTutorial(),makeGUI(),generateData(!0),reset(!0),hideControls()},{"./dataset":1,"./heatmap":2,"./linechart":3,"./nn":4,"./state":6}],6:[function(require,module,exports){"use strict";function getKeyFromValue(obj,value){for(var key in obj)if(obj[key]===value)return key}function endsWith(s,suffix){return s.substr(-suffix.length)===suffix}function getHideProps(obj){var result=[];for(var prop in obj)endsWith(prop,"_hide")&&result.push(prop);return result}var nn=require("./nn"),dataset=require("./dataset");exports.activations={relu:nn.Activations.RELU,tanh:nn.Activations.TANH,sigmoid:nn.Activations.SIGMOID,linear:nn.Activations.LINEAR},exports.regularizations={none:null,L1:nn.RegularizationFunction.L1,L2:nn.RegularizationFunction.L2},exports.datasets={circle:dataset.classifyCircleData,xor:dataset.classifyXORData,gauss:dataset.classifyTwoGaussData,spiral:dataset.classifySpiralData},exports.regDatasets={"reg-plane":dataset.regressPlane,"reg-gauss":dataset.regressGaussian},exports.getKeyFromValue=getKeyFromValue;var Type;!function(Type){Type[Type.STRING=0]="STRING",Type[Type.NUMBER=1]="NUMBER",Type[Type.ARRAY_NUMBER=2]="ARRAY_NUMBER",Type[Type.ARRAY_STRING=3]="ARRAY_STRING",Type[Type.BOOLEAN=4]="BOOLEAN",Type[Type.OBJECT=5]="OBJECT"}(Type=exports.Type||(exports.Type={}));var Problem;!function(Problem){Problem[Problem.CLASSIFICATION=0]="CLASSIFICATION",Problem[Problem.REGRESSION=1]="REGRESSION"}(Problem=exports.Problem||(exports.Problem={})),exports.problems={classification:Problem.CLASSIFICATION,regression:Problem.REGRESSION};var State=function(){function State(){this.learningRate=.03,this.regularizationRate=0,this.showTestData=!1,this.noise=0,this.batchSize=10,this.discretize=!1,this.tutorial=null,this.percTrainData=50,this.activation=nn.Activations.TANH,this.regularization=null,this.problem=Problem.CLASSIFICATION,this.initZero=!1,this.hideText=!1,this.collectStats=!1,this.numHiddenLayers=1,this.hiddenLayerControls=[],this.networkShape=[4,2],this.x=!0,this.y=!0,this.xTimesY=!1,this.xSquared=!1,this.ySquared=!1,this.cosX=!1,this.sinX=!1,this.cosY=!1,this.sinY=!1,this.dataset=dataset.classifyCircleData,this.regDataset=dataset.regressPlane}return State.deserializeState=function(){function hasKey(name){return name in map&&null!=map[name]&&""!==map[name].trim()}function parseArray(value){return""===value.trim()?[]:value.split(",")}for(var map={},_i=0,_a=window.location.hash.slice(1).split("&");_i<_a.length;_i++){var keyvalue=_a[_i],_b=keyvalue.split("="),name_1=_b[0],value=_b[1];map[name_1]=value}var state=new State;return State.PROPS.forEach(function(_a){var name=_a.name,type=_a.type,keyMap=_a.keyMap;switch(type){case Type.OBJECT:if(null==keyMap)throw Error("A key-value map must be provided for state variables of type Object");hasKey(name)&&map[name]in keyMap&&(state[name]=keyMap[map[name]]);break;case Type.NUMBER:hasKey(name)&&(state[name]=+map[name]);break;case Type.STRING:hasKey(name)&&(state[name]=map[name]);break;case Type.BOOLEAN:hasKey(name)&&(state[name]="false"!==map[name]);break;case Type.ARRAY_NUMBER:name in map&&(state[name]=parseArray(map[name]).map(Number));break;case Type.ARRAY_STRING:name in map&&(state[name]=parseArray(map[name]));break;default:throw Error("Encountered an unknown type for a state variable")}}),getHideProps(map).forEach(function(prop){state[prop]="true"===map[prop]}),state.numHiddenLayers=state.networkShape.length,null==state.seed&&(state.seed=Math.random().toFixed(5)),Math.seedrandom(state.seed),state},State.prototype.serialize=function(){var _this=this,props=[];State.PROPS.forEach(function(_a){var name=_a.name,type=_a.type,keyMap=_a.keyMap,value=_this[name];null!=value&&(type===Type.OBJECT?value=getKeyFromValue(keyMap,value):type!==Type.ARRAY_NUMBER&&type!==Type.ARRAY_STRING||(value=value.join(",")),props.push(name+"="+value))}),getHideProps(this).forEach(function(prop){props.push(prop+"="+_this[prop])}),window.location.hash=props.join("&")},State.prototype.getHiddenProps=function(){var result=[];for(var prop in this)endsWith(prop,"_hide")&&"true"===String(this[prop])&&result.push(prop.replace("_hide",""));return result},State.prototype.setHideProperty=function(name,hidden){this[name+"_hide"]=hidden},State}();State.PROPS=[{name:"activation",type:Type.OBJECT,keyMap:exports.activations},{name:"regularization",type:Type.OBJECT,keyMap:exports.regularizations},{name:"batchSize",type:Type.NUMBER},{name:"dataset",type:Type.OBJECT,keyMap:exports.datasets},{name:"regDataset",type:Type.OBJECT,keyMap:exports.regDatasets},{name:"learningRate",type:Type.NUMBER},{name:"regularizationRate",type:Type.NUMBER},{name:"noise",type:Type.NUMBER},{name:"networkShape",type:Type.ARRAY_NUMBER},{name:"seed",type:Type.STRING},{name:"showTestData",type:Type.BOOLEAN},{name:"discretize",type:Type.BOOLEAN},{name:"percTrainData",type:Type.NUMBER},{name:"x",type:Type.BOOLEAN},{name:"y",type:Type.BOOLEAN},{name:"xTimesY",type:Type.BOOLEAN},{name:"xSquared",type:Type.BOOLEAN},{name:"ySquared",type:Type.BOOLEAN},{name:"cosX",type:Type.BOOLEAN},{name:"sinX",type:Type.BOOLEAN},{name:"cosY",type:Type.BOOLEAN},{name:"sinY",type:Type.BOOLEAN},{name:"collectStats",type:Type.BOOLEAN},{name:"tutorial",type:Type.STRING},{name:"problem",type:Type.OBJECT,keyMap:exports.problems},{name:"initZero",type:Type.BOOLEAN},{name:"hideText",type:Type.BOOLEAN}],exports.State=State},{"./dataset":1,"./nn":4}]},{},[5]);
favicon.png ADDED
index.html CHANGED
@@ -42,7 +42,7 @@ limitations under the License.
42
  </head>
43
  <body>
44
  <!-- GitHub link -->
45
- <a class="github-link" href="https://github.com/tensorflow/playground" title="Source on GitHub" target="_blank">
46
  <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" viewBox="0 0 60.5 60.5" width="60" height="60">
47
  <polygon class="bg" points="60.5,60.5 0,0 60.5,0 "/>
48
  <path class="icon" d="M43.1,5.8c-6.6,0-12,5.4-12,12c0,5.3,3.4,9.8,8.2,11.4c0.6,0.1,0.8-0.3,0.8-0.6c0-0.3,0-1,0-2c-3.3,0.7-4-1.6-4-1.6c-0.5-1.4-1.3-1.8-1.3-1.8c-1.1-0.7,0.1-0.7,0.1-0.7c1.2,0.1,1.8,1.2,1.8,1.2c1.1,1.8,2.8,1.3,3.5,1c0.1-0.8,0.4-1.3,0.8-1.6c-2.7-0.3-5.5-1.3-5.5-5.9c0-1.3,0.5-2.4,1.2-3.2c-0.1-0.3-0.5-1.5,0.1-3.2c0,0,1-0.3,3.3,1.2c1-0.3,2-0.4,3-0.4c1,0,2,0.1,3,0.4c2.3-1.6,3.3-1.2,3.3-1.2c0.7,1.7,0.2,2.9,0.1,3.2c0.8,0.8,1.2,1.9,1.2,3.2c0,4.6-2.8,5.6-5.5,5.9c0.4,0.4,0.8,1.1,0.8,2.2c0,1.6,0,2.9,0,3.3c0,0.3,0.2,0.7,0.8,0.6c4.8-1.6,8.2-6.1,8.2-11.4C55.1,11.2,49.7,5.8,43.1,5.8z"/>
 
42
  </head>
43
  <body>
44
  <!-- GitHub link -->
45
+ <a class="github-link" href="https://github.com/tensorflow/playground" title="Source on GitHub">
46
  <svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" viewBox="0 0 60.5 60.5" width="60" height="60">
47
  <polygon class="bg" points="60.5,60.5 0,0 60.5,0 "/>
48
  <path class="icon" d="M43.1,5.8c-6.6,0-12,5.4-12,12c0,5.3,3.4,9.8,8.2,11.4c0.6,0.1,0.8-0.3,0.8-0.6c0-0.3,0-1,0-2c-3.3,0.7-4-1.6-4-1.6c-0.5-1.4-1.3-1.8-1.3-1.8c-1.1-0.7,0.1-0.7,0.1-0.7c1.2,0.1,1.8,1.2,1.8,1.2c1.1,1.8,2.8,1.3,3.5,1c0.1-0.8,0.4-1.3,0.8-1.6c-2.7-0.3-5.5-1.3-5.5-5.9c0-1.3,0.5-2.4,1.2-3.2c-0.1-0.3-0.5-1.5,0.1-3.2c0,0,1-0.3,3.3,1.2c1-0.3,2-0.4,3-0.4c1,0,2,0.1,3,0.4c2.3-1.6,3.3-1.2,3.3-1.2c0.7,1.7,0.2,2.9,0.1,3.2c0.8,0.8,1.2,1.9,1.2,3.2c0,4.6-2.8,5.6-5.5,5.9c0.4,0.4,0.8,1.1,0.8,2.2c0,1.6,0,2.9,0,3.3c0,0.3,0.2,0.7,0.8,0.6c4.8-1.6,8.2-6.1,8.2-11.4C55.1,11.2,49.7,5.8,43.1,5.8z"/>
lib.js CHANGED
The diff for this file is too large to render. See raw diff
 
preview.png ADDED