This repository contains clean, modular implementations of two foundational Deep Reinforcement Learning algorithms:
- Proximal Policy Optimization (PPO)
- Soft Actor-Critic (SAC)
Both are implemented using PyTorch and are suited for continuous action spaces.
This repo is designed for:
- 🔬 Researchers
- 📚 Students
- 🧑💻 Developers building DRL pipelines
Keywords: Soft Actor-Critic, PPO, Reinforcement Learning, PyTorch, DRL, RL, Actor Critic, Continuous Control, OpenAI Gym, Off-policy, On-policy, Deep RL, SOTA
- 🧱 Modular structure with reusable components (Actor, Critic, Memory)
- 🧮 PPO implementation with GAE, clipping, entropy regularisation
- 🔁 SAC with automatic entropy tuning and twin Q-networks
- 🔍 Logging of rewards and action distributions
- ⚙️ Easy to integrate into any RL environment
git clone https://github.com/your-username/ppo-sac-pytorch.git
cd ppo-sac-pytorch
pip install -r requirements.txt