30:T45c,function convertToTensor(data) { // Wrapping these calculations in a tidy will dispose any // intermediate tensors. return tf.tidy(() => { // Step 1. Shuffle the data tf.util.shuffle(data); // Step 2. Convert data to Tensor const inputs = data.map((d) => d.horsepower); const labels = data.map((d) => d.mpg); const inputTensor = tf.tensor2d(inputs, [inputs.length, 1]); const labelTensor = tf.tensor2d(labels, [labels.length, 1]); //Step 3. Normalize the data to the range 0 - 1 using min-max scaling const inputMax = inputTensor.max(); const inputMin = inputTensor.min(); const labelMax = labelTensor.max(); const labelMin = labelTensor.min(); const normalizedInputs = inputTensor .sub(inputMin) .div(inputMax.sub(inputMin)); const normalizedLabels = labelTensor .sub(labelMin) .div(labelMax.sub(labelMin)); return { inputs: normalizedInputs, labels: normalizedLabels, // Return the min/max bounds so we can use them later. inputMax, inputMin, labelMax, labelMin }; }); }31:T46e,function testModel(model, inputData, normalizationData) { const { inputMax, inputMin, labelMin, labelMax } = normalizationData; // Generate predictions for a uniform range of numbers between 0 and 1; // We un-normalize the data by doing the inverse of the min-max scaling // that we did earlier. const [xs, preds] = tf.tidy(() => { const xs = tf.linspace(0, 1, 100); const preds = model.predict(xs.reshape([100, 1])); const unNormXs = xs.mul(inputMax.sub(inputMin)).add(inputMin); const unNormPreds = preds.mul(labelMax.sub(labelMin)).add(labelMin); // Un-normalize the data return [unNormXs.dataSync(), unNormPreds.dataSync()]; }); const predictedPoints = Array.from(xs).map((val, i) => { return { x: val, y: preds[i] }; }); const originalPoints = inputData.map((d) => ({ x: d.horsepower, y: d.mpg })); tfvis.render.scatterplot( { name: "Model Predictions vs Original Data" }, { values: [originalPoints, predictedPoints], series: ["original", "predicted"] }, { xLabel: "Horsepower", yLabel: "MPG", height: 300 } ); }32:[] 2e:{"title":"Tensor Flow","cells":[{"id":0,"value":"# Tensor Flow\n\nExploring tensorflowJS v3⁺, with the help of codelabs.","pinned":false,"hidden":false,"mode":"md"},{"id":83,"value":"## CodeLab regression\nStarting with tfjs-training-regression.","pinned":false,"hidden":false,"mode":"md"},{"id":161,"value":"### Get car data","pinned":false,"hidden":false,"mode":"md"},{"id":168,"value":"cars = tfEpi.getCars()","pinned":false,"hidden":false,"mode":"ojs"},{"id":174,"value":"Plot.plot({\n marks: [Plot.dot(cars, { x: \"mpg\", y: \"horsepower\" })]\n})","pinned":false,"hidden":false,"mode":"ojs"},{"id":181,"value":"async function run() {\n // Load and plot the original input data that we are going to train on.\n const data = await tfEpi.getCars();\n const values = data.map((d) => ({\n x: d.horsepower,\n y: d.mpg\n }));\n\n tfvis.render.scatterplot(\n { name: \"Horsepower v MPG\" },\n { values },\n {\n xLabel: \"Horsepower\",\n yLabel: \"MPG\",\n height: 300\n }\n );\n\n // More code will be added below\n // Convert the data to a form we can use for training.\n const tensorData = convertToTensor(data);\n const { inputs, labels } = tensorData;\n const model = createModel();\n // Train the model\n await trainModel(model, inputs, labels);\n console.log(\"Done Training\");\n // Make some predictions using the model and compare them to the\n // original data\n testModel(model, data, tensorData);\n return model;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":185,"value":"model = run()","pinned":false,"hidden":false,"mode":"ojs"},{"id":320,"value":"model.getConfig()","pinned":false,"hidden":false,"mode":"ojs"},{"id":347,"value":"model.save(\"localstorage://my-model-1\")","pinned":true,"hidden":false,"mode":"ojs"},{"id":356,"value":"model.save(\"\")","pinned":true,"hidden":false,"mode":"ojs"},{"id":187,"value":"function createModel() {\n // Create a sequential model\n const model = tf.sequential();\n\n // Add a single input layer\n model.add(tf.layers.dense({ inputShape: [1], units: 1, useBias: true }));\n\n // hidden layer\n model.add(\n tf.layers.dense({ units: 50, activation: \"sigmoid\", useBias: true })\n );\n model.add(\n tf.layers.dense({ units: 50, activation: \"sigmoid\", useBias: true })\n );\n model.add(\n tf.layers.dense({ units: 50, activation: \"sigmoid\", useBias: true })\n );\n\n // Add an output layer\n model.add(tf.layers.dense({ units: 1, useBias: true }));\n\n return model;\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":191,"value":"$30","pinned":false,"hidden":false,"mode":"ojs"},{"id":293,"value":"async function trainModel(model, inputs, labels) {\n // Prepare the model for training.\n model.compile({\n optimizer: tf.train.adam(),\n loss: tf.losses.meanSquaredError,\n metrics: [\"mse\"]\n });\n\n const batchSize = 32;\n const epochs = 100;\n\n return await model.fit(inputs, labels, {\n batchSize,\n epochs,\n shuffle: true,\n callbacks: tfvis.show.fitCallbacks(\n { name: \"Training Performance\" },\n [\"loss\", \"mse\"],\n { height: 200, callbacks: [\"onEpochEnd\"] }\n )\n });\n}","pinned":false,"hidden":false,"mode":"ojs"},{"id":298,"value":"$31","pinned":false,"hidden":false,"mode":"ojs"},{"id":7,"value":"## Methods","pinned":false,"hidden":false,"mode":"md"},{"id":17,"value":"### Core libraries\nRegular tensor flow, tf , and the visualization sidekick, tfvis , and epiSphere's auxiliary tfEpi library.\n`","pinned":false,"hidden":false,"mode":"md"},{"id":9,"value":"tf = require(\"@tensorflow/tfjs\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":58,"value":"tfvis = require(\"@tensorflow/tfjs-vis\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":14,"value":"tfEpi = require(\"https://episphere.github.io/tfjs/tfEpi.js\")\n//tfEpi = require(\"http://localhost:8000/tfjs/tfEpi.js\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":220,"value":"stats = require(\"simple-statistics\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":245,"value":"Plotly = require(\"https://cdn.plot.ly/plotly-2.4.2.min.js\")","pinned":false,"hidden":false,"mode":"ojs"},{"id":193,"value":"## MIS\nMiscellaneous operations","pinned":false,"hidden":false,"mode":"md"},{"id":202,"value":"x = tf.tensor2d([1, 2, 3, 4, 5, 6, 7], [7, 1])","pinned":true,"hidden":false,"mode":"ojs"},{"id":213,"value":"inputTensor = tf.tensor2d(\n cars.map((d) => d.horsepower),\n [cars.length, 1]\n)","pinned":true,"hidden":false,"mode":"ojs"},{"id":216,"value":"labelTensor = tf.tensor2d(\n cars.map((d) => d.mpg),\n [cars.length, 1]\n)","pinned":true,"hidden":false,"mode":"ojs"},{"id":218,"value":"inputMax = inputTensor.max()","pinned":true,"hidden":false,"mode":"ojs"},{"id":238,"value":"inputMin = inputTensor.min()","pinned":true,"hidden":false,"mode":"ojs"},{"id":223,"value":"[stats.max(cars.map((d) => d.horsepower)), stats.max(cars.map((d) => d.mpg))]","pinned":true,"hidden":false,"mode":"ojs"},{"id":235,"value":"normalizedInputs = inputTensor.sub(inputMin).div(inputMax.sub(inputMin))","pinned":true,"hidden":false,"mode":"ojs"},{"id":242,"value":"normalizedInputs.data()","pinned":true,"hidden":false,"mode":"ojs"},{"id":248,"value":"Plotly.newPlot(DOM.element(\"div\"), [\n {\n x: (await normalizedInputs.array()).map((d) => d[0]),\n type: \"histogram\"\n }\n])","pinned":true,"hidden":false,"mode":"ojs"},{"id":261,"value":"arr = (await normalizedInputs.array()).map((d) => d[0])","pinned":true,"hidden":false,"mode":"ojs"},{"id":275,"value":"arr.slice(0, 4).map","pinned":true,"hidden":false,"mode":"ojs"}],"files":"$W32","stdlib":"1","titling":"auto"}