TF2RL provides implementations of deep reinforcement learning algorithms including policy gradient methods (VPG, PPO), value-based methods (DQN, DDQN, Prioritized DQN, Dueling DQN, Distributional DQN, Noisy DQN), and actor-critic methods (DDPG, TD3, SAC). It supports both discrete and continuous action spaces and includes features like GAE for advantage estimation and Ape-X for distributed training.