all-rl-algorithms

What it is

This repository is a collection of Python implementations of various Reinforcement Learning (RL) algorithms. The primary goal is educational, to understand how these algorithms work under the hood. It includes a comprehensive cheat sheet (cheatsheet.md) and emphasizes readability over performance. The project lists 18 algorithm notebooks as part of updates.

How it works

Each algorithm has its own Jupyter Notebook file (.ipynb) with detailed explanations and implementations. The README outlines a learning guide that directs users from basics to advanced methods, and mentions that the A3C implementation uses a separate script (a3c_training.py) due to multiprocessing in notebooks. The repository uses basic libraries (NumPy, Matplotlib, PyTorch).

Getting started

Follow installation steps:

# Clone the repository
git clone https://github.com/fareedkhan-dev/all-rl-algorithms.git
cd all-rl-algorithms

# Create a virtual environment (using uv)
uv init
uv venv

Activate:

# Windows
.venv\Scripts\activate

# macOS / Linux
source .venv/bin/activate

Install dependencies:

uv add -r requirements.txt

Multiprocessing note for A3C:

# Run a3c_training.py from terminal

If uv is not installed:

pip install uv

Recent releases

• There are no releases listed in the repository metadata for latest version.

Traction

• Stars: 1889

Behind the repo

• Not applicable as no startup/company link is provided in the facts.

Caveats

• License: MIT
• Created: 2025-03-30
• Last push: 2025-08-29
• Language: Jupyter Notebook
• Open issues: 1
• Note: This repository is described as not a performance-optimized library; emphasis is on educational clarity. Some notebooks (e.g., HAC) may contain bugs or incomplete implementations according to the README.