Apollo DQN — reproducible DQN variants for LunarLander-v3 with a Residual MLP backbone, PER-lite replay, strict evaluation pipeline and human baselines.
reinforcement-learning neural-network deep-reinforcement-learning openai-gym q-learning pytorch dqn mlp policy-evaluation residual-networks gymnasium model-evaluation deep-q-learning double-dqn dueling-dqn prioritized-experience-replay lunar-lander experience-replay dueling-ddqn lunarlander-v3
-
Updated
Mar 3, 2026 - Python