MNIST MLP

Interactive Handwritten Digit Recognition

Loading model...

Draw a Digit

Draw a digit (0-9) on the 28×28 canvas. The model predicts in real-time!

Predictions

Digit 00.0%
Digit 10.0%
Digit 20.0%
Digit 30.0%
Digit 40.0%
Digit 50.0%
Digit 60.0%
Digit 70.0%
Digit 80.0%
Digit 90.0%

About This Model

This is a multilayer perceptron (MLP) neural network trained on the MNIST dataset. The model runs entirely in your browser using ONNX Runtime Web, providing real-time predictions without any server communication. Draw a digit and watch the neural network classify it instantly!

Model Architecture

Input Layer

28 × 28 = 784 pixels

Hidden Layer 1

50 neurons + ReLU + Dropout

Hidden Layer 2

50 neurons + ReLU + Dropout

Output Layer

10 classes (0-9)

Technologies

PyTorchONNXONNX Runtime WebNext.jsTypeScriptCanvas API