https://github.com/suquark/inspirenn
https://github.com/suquark/inspirenn
convnetjs deep-learning neural-network presentation
Last synced: 10 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/suquark/inspirenn
- Owner: suquark
- License: mit
- Created: 2017-01-23T15:05:09.000Z (over 9 years ago)
- Default Branch: master
- Last Pushed: 2017-03-19T04:40:39.000Z (over 9 years ago)
- Last Synced: 2025-08-08T04:49:34.126Z (11 months ago)
- Topics: convnetjs, deep-learning, neural-network, presentation
- Language: JavaScript
- Size: 3.51 MB
- Stars: 15
- Watchers: 3
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: Readme.md
- License: LICENSE
Awesome Lists containing this project
README
# InspireNN
This project makes neural networks running on browsers. It can be useful in blogging & presentation.
It reimplements [ConvNetJS](https://github.com/karpathy/convnetjs) with modern styles (ECMAScript6, functional programming), making it's code shorter, more readable for beginners and easier to extend.
It's sure that we should never expect a neural network training in the browser doing a big deal, but it's useful for presentation and understanding.
It is still under developing.
## Dependency
* transpiler
`systemjs` + `traceur` is used to load ES6 modules and provide ES6 polyfill for older browsers.
* style
`jquery`, `materializecss`, `d3js` and `three.js` is included in the project for supporting visualization tasks. Though making use of CDNs may be a better idea, but I prefer to keep them local for offline situations.
## Demo
You may start a http server at root of the project and try some of the demos under `demo/` in your browser.
We will soon provide you with online demos.
### Regression
`demo/regression/` is a demo about regression.
## License
MIT