npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

@microduino/tf-knn-classifier

v1.0.3-alpha

Published

KNN Classifier for TensorFlow.js

Downloads

3

Readme

KNN Classifier

This package provides a utility for creating a classifier using the K-Nearest Neighbors algorithm.

This package is different from the other packages in this repository in that it doesn't provide a model with weights, but rather a utility for constructing a KNN model using activations from another model or any other tensors you can associate with a class/label.

You can see example code here.

Usage example

via Script Tag
<html>
  <head>
    <!-- Load TensorFlow.js -->
    <script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs"></script>
    <!-- Load MobileNet -->
    <script src="https://cdn.jsdelivr.net/npm/@tensorflow-models/mobilenet"></script>
    <!-- Load KNN Classifier -->
    <script src="https://cdn.jsdelivr.net/npm/@tensorflow-models/knn-classifier"></script>
 </head>

  <body>
    <img id='class0' src='/images/class0.jpg '/>
    <img id='class1' src='/images/class1.jpg '/>
    <img id='test' src='/images/test.jpg '/>
  </body>
  <!-- Place your code in the script tag below. You can also use an external .js file -->
  <script>

    const init = async function() {
      // Create the classifier.
      const classifier = knnClassifier.create();

      // Load mobilenet.
      const mobilenetModule = await mobilenet.load();

      // Add MobileNet activations to the model repeatedly for all classes.
      const img0 = tf.browser.fromPixels(document.getElementById('class0'));
      const logits0 = mobilenetModule.infer(img0, 'conv_preds');
      classifier.addExample(logits0, 0);

      const img1 = tf.browser.fromPixels(document.getElementById('class1'));
      const logits1 = mobilenetModule.infer(img1, 'conv_preds');
      classifier.addExample(logits1, 1);

      // Make a prediction.
      const x = tf.browser.fromPixels(document.getElementById('test'));
      const xlogits = mobilenetModule.infer(x, 'conv_preds');
      console.log('Predictions:');
      const result = await classifier.predictClass(xlogits);
      console.log(result);
    }

    init();

  </script>
</html>
via NPM
import * as tf from '@tensorflow/tfjs';
import * as mobilenetModule from '@tensorflow-models/mobilenet';
import * as knnClassifier from '@tensorflow-models/knn-classifier';

// Create the classifier.
const classifier = knnClassifier.create();

// Load mobilenet.
const mobilenet = await mobilenetModule.load();

// Add MobileNet activations to the model repeatedly for all classes.
const img0 = tf.browser.fromPixels(document.getElementById('class0'));
const logits0 = mobilenet.infer(img0, 'conv_preds');
classifier.addExample(logits0, 0);

const img1 = tf.browser.fromPixels(document.getElementById('class1'));
const logits1 = mobilenet.infer(img1, 'conv_preds');
classifier.addExample(logits1, 1);

// Make a prediction.
const x = tf.browser.fromPixels(document.getElementById('test'));
const xlogits = mobilenet.infer(x, 'conv_preds');
console.log('Predictions:');
console.log(classifier.predictClass(xlogits));

API

Creating a classifier

knnClassifier is the module name, which is automatically included when you use the method.

classifier = knnClassifier.create()

Returns a KNNImageClassifier.

Adding examples

classifier.addExample(
  example: tf.Tensor,
  label: number|string
): void;

Args:

  • example: An example to add to the dataset, usually an activation from another model.
  • label: The label (class name) of the example.

Making a prediction

classifier.predictClass(
  input: tf.Tensor,
  k = 3
): Promise<{label: string, classIndex: number, confidences: {[classId: number]: number}}>;

Args:

  • input: An example to make a prediction on, usually an activation from another model.
  • k: The K value to use in K-nearest neighbors. The algorithm will first find the K nearest examples from those it was previously shown, and then choose the class that appears the most as the final prediction for the input example. Defaults to 3. If examples < k, k = examples.

Returns an object where:

  • label: the label (class name) with the most confidence.
  • classIndex: the 0-based index of the class (for backwards compatibility).
  • confidences: maps each label to their confidence score.

Misc

Clear all examples for a class.
classifier.clearClass(label: number|string)

Args:

  • label: The label to clear all examples for.
Clear all examples from all classes
classifier.clearAllClasses()
Get the example count for each class
classifier.getClassExampleCount(): {[label: string]: number}

Returns an object that maps label name to example count for that label.

Get the full dataset, useful for saving state.
classifier.getClassifierDataset(): {[label: string]: Tensor2D}
Set the full dataset, useful for restoring state.
classifier.setClassifierDataset(dataset: {[label: string]: Tensor2D})

Args:

  • dataset: The label dataset matrices map. Can be retrieved from getClassDatsetMatrices. Useful for restoring state.
Get the total number of classes
classifier.getNumClasses(): number
Dispose the classifier and all internal state

Clears up WebGL memory. Useful if you no longer need the classifier in your application.

classifier.dispose()