https://github.com/georgestagg/mnist-classifier-blas-wasm

MNIST Classifier - A WebAssembly BLAS Demo
https://github.com/georgestagg/mnist-classifier-blas-wasm

Last synced: 3 months ago
JSON representation

MNIST Classifier - A WebAssembly BLAS Demo

• Host: GitHub
• URL: https://github.com/georgestagg/mnist-classifier-blas-wasm
• Owner: georgestagg
• License: mit
• Created: 2024-03-05T09:24:35.000Z (11 months ago)
• Default Branch: main
• Last Pushed: 2024-03-08T09:48:07.000Z (11 months ago)
• Last Synced: 2024-10-11T18:25:50.682Z (4 months ago)
• Language: JavaScript
• Homepage: https://georgestagg.github.io/mnist-classifier-blas-wasm/
• Size: 2.96 MB
• Stars: 6
• Watchers: 2
• Forks: 2
• Open Issues: 1
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# MNIST Classifier - A WebAssembly BLAS Demo

A handwritten digit classifier demonstrating BLAS routines running in a web browser using WebAssembly.

A [multilayer perceptron network](https://en.wikipedia.org/wiki/Multilayer_perceptron) is used for the modeling and classification of digits. The pre-training of model weights has been performed ahead of time using [Scikit-learn](http://Scikit-learn.org), and the classification of digits runs interactively using JavaScript and WebAssembly.

The [MNIST database](https://en.wikipedia.org/wiki/MNIST_database) provides the source data used to train the model. Classifier output is shown using [Observable Plot](https://observablehq.com/plot/). 

# Interactive Website

Draw a digit from 0-9 in the box and the classifier will try to label the handwritten digit. The resulting relative probabilities will be shown in a plot on the right.

https://georgestagg.github.io/mnist-classifier-blas-wasm/