https://github.com/trixky/digit_classifier
An online digit classifier using the multi-layer perceptrons (MLP) model.
https://github.com/trixky/digit_classifier
classification-model mlp mnist mnist-dataset multilayer-perceptron tensorflow tensorflowjs
Last synced: 7 months ago
JSON representation
An online digit classifier using the multi-layer perceptrons (MLP) model.
- Host: GitHub
- URL: https://github.com/trixky/digit_classifier
- Owner: trixky
- Created: 2024-07-06T13:28:41.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-07-07T13:39:42.000Z (over 1 year ago)
- Last Synced: 2025-01-23T23:53:33.318Z (9 months ago)
- Topics: classification-model, mlp, mnist, mnist-dataset, multilayer-perceptron, tensorflow, tensorflowjs
- Language: Svelte
- Homepage: https://trixky.github.io/digit_classifier/
- Size: 4.63 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# digit_classifier
An [online](https://trixky.github.io/digit_classifier/) digit classifier using the [multi-layer perceptrons](https://en.wikipedia.org/wiki/Multilayer_perceptron) (MLP) model.
The model is trained on the [MNIST dataset](https://en.wikipedia.org/wiki/MNIST_database) and implemented using the [TensorFlow.js](https://www.tensorflow.org/js) library.
> The model is loaded on the client side and all calculations are made on the device.
## Setup
```bash
npm run install
npm run dev #localhost:5173
```
## Model caracteristics
- dataset: MNIST (10 000 samples)
- input layer: 784 (28x28 pixels)
- hidden layers (2): 32 neurons (ReLU activation) + 16 neurons (ReLU activation)
- output layer: 10 (Softmax activation)
- optimizer: Adam
- loss: categoricalCrossentropy
- metrics: accuracy
- epochs: 50
- batch size: 512