NeatRL
An open-source educational library that implements classic reinforcement learning algorithms (DQN, REINFORCE, PPO, SAC, and others) from scratch. The code is concise and does not depend on high-level RL libraries, making it suitable for studying the internals of the algorithms.
Timeline (last 90 days)
NeatRL is an open-source library that implements reinforcement learning algorithms from scratch, covering algorithms such as DQN, REINFORCE, PPO, and SAC.
NeatRL's algorithm implementations do not depend on high-level reinforcement learning libraries; the code is written from scratch with clear, visible logic.
NeatRL's algorithms are implemented in standard environments such as OpenAI Gym and MuJoCo, covering both discrete-action and continuous-action scenarios.
NeatRL was shared by a Reddit user, and the author reintroduced the library in a video and shared usage experience.
NeatRL's target users are people who want to quickly understand how reinforcement learning algorithms work internally.
NeatRL is part of an "ongoing series," and more algorithms and notes may be added in the future.
All Facts (6)
NeatRL is an open-source library that implements reinforcement learning algorithms from scratch, covering algorithms such as DQN, REINFORCE, PPO, and SAC.
50%UnverifiedNeatRL's algorithm implementations do not depend on high-level reinforcement learning libraries; the code is written from scratch with clear, visible logic.
50%UnverifiedNeatRL's algorithms are implemented in standard environments such as OpenAI Gym and MuJoCo, covering both discrete-action and continuous-action scenarios.
50%UnverifiedNeatRL's target users are people who want to quickly understand how reinforcement learning algorithms work internally.
50%UnverifiedNeatRL was shared by a Reddit user, and the author reintroduced the library in a video and shared usage experience.
50%UnverifiedNeatRL is part of an "ongoing series," and more algorithms and notes may be added in the future.
50%