https://github.com/adiagarwalrock/mnist-digits-web
Draw a digit and watch activations/probabilities update in real time.
https://github.com/adiagarwalrock/mnist-digits-web
digits mnist onnx visualization
Last synced: 2 months ago
JSON representation
Draw a digit and watch activations/probabilities update in real time.
- Host: GitHub
- URL: https://github.com/adiagarwalrock/mnist-digits-web
- Owner: adiagarwalrock
- Created: 2026-02-17T09:36:48.000Z (4 months ago)
- Default Branch: master
- Last Pushed: 2026-02-28T02:27:36.000Z (4 months ago)
- Last Synced: 2026-02-28T08:22:13.590Z (4 months ago)
- Topics: digits, mnist, onnx, visualization
- Language: JavaScript
- Homepage: https://digits.adityaagarwal.me/
- Size: 2.06 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# MNIST ONNX Draw Visualizer
Pure frontend app for MNIST digit drawing + live ONNX inference.
Created by **Aditya Agarwal**
Website: `https://www.adityaagarwal.me`
GitHub: `https://github.com/adiagarwalrock`
Repo: `https://github.com/adiagarwalrock/mnist-digits-web`
## Features
- Draw a digit on a square canvas.
- Real-time prediction from `web/model/model.onnx`.
- Output bars for `0-9` probabilities.
- Network visualization with distinct layer sizes:
- Input: derived from input feature size.
- Hidden: computed as a different count (not equal to input/output).
- Output: 10 nodes.
## Local Run (No Python)
Use any static server. Example:
```bash
npm install
npm run preview
```
Then open the printed local URL.
## Deploy to Cloudflare Workers
Files already prepared:
- `wrangler.toml`
- `worker.js`
Commands:
```bash
npm install
npm run deploy:cf
```
For local worker preview:
```bash
npm run dev:cf
```
## Deploy to Vercel
Files already prepared:
- `vercel.json`
Commands:
```bash
npm install
npm run deploy:vercel
```
For local Vercel preview:
```bash
npm run dev:vercel
```
## Model Path
- Default model URL: `./model/model.onnx`
- File location in repo: `web/model/model.onnx`