evelinag/explorable-reinforcement-learning

Explorable explanation of basic reinforcement learning algorithms

Explorable reinforcement learning

Work in progress

Attempt at creating explorable explanation of basic reinforcement learning algorithms.

---

Based on: Fable Minimal App

This is a simple Fable app including an Elmish counter with as little configuration as possible. If you want to see a more complex app including commonly used F# tools like Paket or Fake, check the Fulma demo.

Requirements

• dotnet SDK 2.1 or higher
• node.js with npm
• An F# editor like Visual Studio, Visual Studio Code with Ionide or JetBrains Rider.

Building and running the app

• Install JS dependencies: npm install
• Start Webpack dev server: npx webpack-dev-server or npm start
• After the first compilation is finished, in your browser open: http://localhost:8080/

Any modification you do to the F# code will be reflected in the web page after saving.

Project structure

npm

JS dependencies are declared in package.json, while package-lock.json is a lock file automatically generated.

Webpack

Webpack is a JS bundler with extensions, like a static dev server that enables hot reloading on code changes. Fable interacts with Webpack through the fable-loader. Configuration for Webpack is defined in the webpack.config.js file. Note this sample only includes basic Webpack configuration for development mode, if you want to see a more comprehensive configuration check the Fable webpack-config-template.

F#

The sample only contains two F# files: the project (.fsproj) and a source file (.fs) in the src folder.

Web assets

The index.html file and other assets like an icon can be found in the public folder.