TensorFlow Training
Train the Model
The examples on the previous pages uses this function to train the model:
async function trainModel(model, inputs, labels, surface) {
const batchSize = 25;
const epochs = 100;
const callbacks = tfvis.show.fitCallbacks(surface, ['loss'], {callbacks:['onEpochEnd']})
return await model.fit(inputs, labels,
{batchSize, epochs, shuffle:true, callbacks:callbacks}
);
}
epochs defines how many iterations (loops) the model will do.
model.fit is the function that runs the loops.
callbacks defines the callback function to call when the model wants to redraw the graphics.
Test the Model
When a model is trained, it is important to test and evaluate it.
We do this by inspecting what the model predicts for a range of different inputs.
But, before we can do that, we have to un-normalize the data:
Un Normalize
let unX = tf.linspace(0, 1, 100);
let unY = model.predict(unX.reshape([100, 1]));
const unNormunX = unX.mul(inputMax.sub(inputMin)).add(inputMin);
const unNormunY = unY.mul(labelMax.sub(labelMin)).add(labelMin);
unX = unNormunX.dataSync();
unY = unNormunY.dataSync();
Then we can look at the result:
Plot the Result
const predicted = Array.from(unX).map((val, i) => {
return {x: val, y: unY[i]}
});
// Plot the Result
tfPlot([values, predicted], surface1)
