Procgen

Procgen Benchmark This dataset contains expert trajectories generated by a PPO reinforcement learning agent trained on each of the 16 procedurally-generated gym environments from the Procgen Benchmark. The environments were created on distribution_mode=easy and with unlimited levels. Disclaimer: This is not an official

Papers using Procgen (15)

ULTHO: Ultra-Lightweight yet Efficient Hyperparameter Optimization in Deep Reinforcement Learning2025 · 1 citesSparse Masked Attention Policies for Reliable Generalization2026Understanding and Improving Hyperbolic Deep Reinforcement Learning2025Synthetic Data is Sufficient for Zero-Shot Visual Generalization from Offline Data2025Impoola: The Power of Average Pooling for Image-Based Deep Reinforcement Learning2025Adaptive Data Exploitation in Deep Reinforcement Learning2025A Dual-Agent Adversarial Framework for Robust Generalization in Deep Reinforcement Learning2025A Study of Plasticity Loss in On-Policy Deep Reinforcement Learning2024 · 1 citesExplore-Go: Leveraging Exploration for Generalisation in Deep Reinforcement Learning2024 · 1 citesDo Agents Dream of Electric Sheep?: Improving Generalization in Reinforcement Learning through Generative Learning2024Fast TRAC: A Parameter-Free Optimizer for Lifelong Reinforcement Learning2024PufferLib: Making Reinforcement Learning Libraries and Environments Play Nice2024SiT: Symmetry-Invariant Transformers for Generalisation in Reinforcement Learning2024Maximum Entropy On-Policy Actor-Critic via Entropy Advantage Estimation2024Provable Domain Adaptation for Offline Reinforcement Learning with Limited Samples2024