finds.dev

wangshusen/DRL

Deep Reinforcement Learning

View on GitHub

★ 4,669NOASSERTIONupdated Dec 2022

// the take

A lecture slide deck and notes collection for deep reinforcement learning, covering Q-learning, policy gradient, actor-critic, DQN variants, multi-agent RL, and continuous action spaces. It's course material from a Chinese university professor, not a library or framework — you're getting PDFs and a lecture note, nothing you can run. Target audience is students or self-learners who want a structured academic introduction to DRL.

The topic coverage is solid for an introductory course: TD learning, DQN improvements (double DQN, dueling networks, prioritized replay), policy gradient with baseline, and A2C are all there. The progression from basics through advanced topics is logical and well-sequenced. Slides are paired with YouTube lectures for each topic, so you're not left decoding static PDFs alone. The multi-agent section is a genuine differentiator — most beginner DRL resources skip MARL entirely.

There is zero code in this repo. No implementations, no Jupyter notebooks, no gym environments — just PDFs. Last commit was December 2022, so PPO, SAC, and anything from the post-ChatGPT RL renaissance (RLHF, GRPO) are absent. The lectures are in Chinese, which cuts off a large part of the audience the slides alone would serve. Imitation learning and partial observation sections are listed but marked incomplete, so the curriculum has visible gaps.

We read the README, the directory tree and a key source file in July 2026. Nobody pays to be listed. How we judge →

// at a glance

Stars
4,669
Forks
679
License
NOASSERTION
Last commit
10 Dec 2022

// nearby

If DRL isn't quite it.

All repos →