DeepRL

Implementations of core Deep Reinforcement Learning (DRL) algorithms in PyTorch for Udacity Deep Reinforcement Learning course.

Algorithms

• DQN — value-based learning for discrete actions
• DDPG — actor–critic for continuous actions
• MADDPG — multi-agent extension of DDPG

Project structure

DeepRL/
├─ DQN/         # Deep Q-Network implementations & notebooks
├─ DDPG/        # Deep Deterministic Policy Gradient
├─ MADDPG/      # Multi-Agent DDPG
├─ utils.py     # shared helpers/utilities
├─ **init**.py
└─ .gitignore

Notes

• Most experiments are notebook-driven; check each folder for environment choice, hyperparameters, and logging.
• Check each algorithm's README.md file for the env setup guidelines.
• Open the notebooks in the algorithm folder you’re interested in (e.g., DQN/, DDPG/, MADDPG/) and run the cells.
• Issues and PRs for improvements, new environments, and benchmarks are welcome.