This repository documents my journey through the Hugging Face Deep Reinforcement Learning Course. It contains my progress, notes, implementations, notebooks, environment setups, and the agents I've trained to master various tasks.
- Understand the theoretical foundations of Reinforcement Learning.
- Learn to use libraries like Stable Baselines3, Gymnasium, and RL Baselines3 Zoo.
- Train agents to play games (Lunar Lander, Atari) and solve robotics tasks.
- Share my models on the Hugging Face Hub.
| Unit | Topic | Status | Model Hub Link | Notes Link |
|---|---|---|---|---|
| 0 | Introduction & Setup | ✅ | - | Notes |
| 1 | Intro to Deep RL (LunarLander-v3) |
✅ | ppo-LunarLander-v3 | Notes |
| 1b | Bonus: Huggy-the-Dog |
✅ | ppo-Huggy | Notes |
| 2 | Q-Learning (FrozenLake-v1 & Taxi-v3) |
✅ | q-FrozenLake-v1-4x4-noSlippery, q-Taxi-v3 |
Notes |
| 3 | Deep Q-Learning (Atari Space Invaders) |
✅ | dqn-SpaceInvadersNoFrameskip-v4 | Notes |
| 3b | Bonus: Optuna Hyperparameter Tuning |
🏗️ | [Link] | Notes |
| 4 | Policy Gradients (CartPole-v1) |
✅ | Reinforce-CartPole-v2, Pixelcopter-PLE-v0 |
Notes |
| 5 | Unity ML-Agents | ✅ | ppo-SnowballTarget ppo-Pyramids |
Notes |
| 6 | Actor-Critic Methods (Robotics) | ✅ | a2c-PandaReachDense-v3 | Notes |
| 7 | Multi-Agent RL & AI vs AI (Soccer) | ✅ | poca-SoccerTwos | Notes |
| 8 | PPO Part 1: Theory & Implementation | ✅ | ppo-LunarLander-v3-02 | Notes |
| 8b | PPO Part 2: VizDoom |
✅ | vizdoom | Notes |
| 9 | Bonus: Advanced Topics in RL | 🏗️ | - | Notes |
| 10 | Bonus: Imitation Learning (Godot) |
🏗️ | - | Notes |
Almost done ...
- Hugging Face Deep RL Course: The main course curriculum.
- Reinforcement Learning: An Introduction (Sutton & Barto): The "Bible" of RL (essential for deep theoretical understanding).
- Clone the repo:
git clone https://github.com/chizkidd/huggingface-deep-RL-course.git cd huggingface-deep-RL-course