TensorflowJS - Template

This is a template for developing interactive machine-learning courses with LiaScript and TensorFlow.js.

Try it on LiaScript:

https://liascript.github.io/course/?https://raw.githubusercontent.com/liaTemplates/tensorflowjs/master/README.md

See the project on Github:

https://github.com/liaTemplates/tensorflowjs

                     --{{1}}--

There are three ways to use this template. The easiest way is to use the import statement and the url of the raw text-file of the master branch or any other branch or version. But you can also copy the required functionionality directly into the header of your Markdown document, see therefor the last slide. And of course, you could also clone this project and change it, as you wish.

                       {{1}}

1. Load the macros via

   import: https://raw.githubusercontent.com/liaTemplates/tensorflowjs/master/README.md

2. Copy the definitions into your Project

3. Clone this repository on GitHub

@TF.eval

                     --{{0}}--

Add the macro @TF.eval to the end of every JavaScript code-block that runs some TensorFlow code and that you want to make editable in LiaScript. The given code gets evaluated asynchroniously and the result is shown in a console below.

// Notice there is no 'import' statement. 'tf' is available on the index-page
// because of the script tag above.

// Define a model for linear regression.
const model = tf.sequential();
model.add(tf.layers.dense({units: 1, inputShape: [1]}));

// Prepare the model for training: Specify the loss and the optimizer.
model.compile({loss: 'meanSquaredError', optimizer: 'sgd'});

// Generate some synthetic data for training.
const xs = tf.tensor2d([1, 2, 3, 4], [4, 1]);
const ys = tf.tensor2d([1, 3, 5, 7], [4, 1]);

// Train the model using the data.
model.fit(xs, ys, {epochs: 10}).then(() => {
  // Use the model to do inference on a data point the model hasn't seen before:
  // Open the browser devtools to see the output
  model.predict(tf.tensor2d([5], [1, 1])).print();
});

@TF.eval

Implementation

                     --{{0}}--

The code shows how the macro @TF.eval is implemented. The script command at the top loads the TensorFlowJS javascript library and onload is used to define function reportError at the initialization phase.

script:   https://cdn.jsdelivr.net/npm/@tensorflow/tfjs@0.13.3/dist/tf.min.js

@onload
window.reportError = function(error) {
  let line = getLineNumber(error);
  let details = [];
  let msg = "An error occured";

  if (line) {
    details = [[{ row : line-1,
               column : 0,
                 text : error.message,
                 type : "error" }]];

    msg += " on line " + line;
  }
  return [ msg + "\n" + error.message, details ];
};
@end

@TF.eval
<script>
async function eee(code) {
  let oldLog = window.console.log;

  window.console.log = console.log;

  try {
    const evalString = '(async function runner() { try { ' + code + '} catch (e) { reportError(e) } })()';

    await eval(evalString).catch(function(e) {
      window.console.log = oldLog;
      let [msg, details] = reportError(e);
      send.lia(msg, details, false);
      send.lia("LIA: stop");
    });
  }
  catch(e) {
    window.console.log = oldLog;
    let [msg, details] = reportError(e);
    send.lia(msg, details, false);
    send.lia("LIA: stop");
  }
  send.lia("LIA: stop");

};
setTimeout(function(e){ eee(`@input`+"\n") }, 10);
"LIA: wait";
</script>
@end

                     --{{1}}--

If you want to minimize loading effort in your LiaScript project, you can also copy this code and paste it into your main comment header, see the code in the raw file of this document.

                       {{1}}

https://raw.githubusercontent.com/liaTemplates/tensorflowjs/master/README.md