Super Mario Bros Stable Baselines3, Now, I’m trying to use stable-baselines3 JAX (SBX) in the same environment but encounter this The superior performance of these networks is verified by engaging three types of classic games, Tic-Tac-Toe, Super Mario Bros. Nov 13, 2025 · The dependency chain in this project requires careful version management to ensure compatibility between gym, gym-super-mario-bros, and stable_baselines3. with Stable-Baseline3 PPO ¶ Super Mario Bros is a well-known video game title developed and published by Nintendo in the 1980s. The project uses stable_baselines3 as the RL framework and gym-super-mario-bros as the game environment. The core logic is currently centralized in mario. My implementation of an RL model to play the NES Super Mario Bros using Stable-Baselines3 (SB3). . The project uses Python 3. This was not an easy task as I only had one course in machine learning and had to learn reinforcement learning from scratch. This script loads a trained PPO model and executes it in the Super Mario Bros environment, rendering gameplay in real-time and displaying reward information. By meticulously addressing errors stemming from version disparities, we provide a systematic guide to navigate through the implementation process successfully. Train an AI to play Super Mario Bros! Uses PPO with a CNN policy to learn from raw pixel inputs. For information about running training after installation, see Training Your First Model. I’m thrilled to share that my latest research paper, "Mastering Super Mario Bros. 0. py10-20 which applies the full wrapper chain. Super Mario Bros. Built with PyTorch & Stable-Baselines3. 0 serves as the compatibility bridge, carefully chosen to satisfy both the older requirements of gym-super-mario-bros and the forward-compatible nature of stable_baselines3. Since gym-retro is in maintenance now and doesn't accept new games, platforms or bug fixes, you can instead submit PRs with new games or features here in stable-retro. , and Car Racing, and achieving the same or even higher levels comparable to human players. Each environment is created via the make_env () function in train. Nov 18, 2025 · A reinforcement learning training/testing example for "Super Mario Bros. A fork of gym-retro ('lets you turn classic video games into Gymnasium environments for reinforcement learning') with additional games, emulators and supported platforms. This research paper tackles the intricate process of implementing Reinforcement Learning (RL) algorithms for training Aug 7, 2024 · Stable Baselines 3 （以下SB3）というライブラリがあるそうなのでそれを使うことにした。 Stable Baselines 3 SB3では強化学習の様々なアルゴリズムが実装されていて、簡単に利用できるようになっている。 Train an AI to play Super Mario Bros! Uses PPO with a CNN policy to learn from raw pixel inputs. : Overcoming Implementation Challenges in Reinforcement Learning with Stable-Baselines3," has been published in RL Baselines3 Zoo provides a collection of pre-trained agents, scripts for training, evaluating agents, tuning hyperparameters, plotting results, and recording videos. We offer a focused approach, emphasizing the utilization of the latest versions of libraries such as OpenAI Gym and Stable-Baselines3 in PyTorch. - ramezaboud/super-mario-rl-agent Nov 13, 2025 · The most critical insight is that gym==0. It is a 2D side-scrolling game, allowing the player to control the main character — Mario. 7, and a two-tier dependency management Nov 13, 2025 · The SubprocVecEnv wrapper from stable_baselines3 runs 8 independent Mario environments in separate processes. In the ever-evolving landscape of artificial intelligence, the application of reinforcement learning (RL) techniques to game playing has emerged as a captivating frontier, showcasing the capacity of intelligent agents to master complex environments and tasks autonomously. Nov 13, 2025 · Purpose and Scope This wiki documents the RL_SuperMario project, a reinforcement learning training system that teaches an agent to play Super Mario Bros using the Proximal Policy Optimization (PPO) algorithm. 8, PyTorch 2. Currently added games on top of gym-retro: Super Mario For one of my project I decided to try and complete the first level of Super Mario Bros using the PPO implementation from Stable Baselines 3 library. 21. - ramezaboud/super-mario-rl-agent Aug 28, 2025 · I previously implemented SAC with stable-baselines3 in a custom Gymnasium environment, and it worked. py, the model evaluation script in the RL_SuperMario project. For detailed dependency version rationale and compatibility constraints, see Dependency Management. As of today (Aug 14 2022) the trained PPO agent completed World 1-1. " based on Stable-Baselines3 (PPO). Nov 13, 2025 · Overview The RL_SuperMario project employs strict version pinning across all dependencies to ensure compatibility between three critical but version-incompatible libraries: gym-super-mario-bros (which requires older gym API), gym (environment interface), and stable_baselines3 (RL algorithms). It is one of the classical game titles that lived through the years and need no explanations. Nov 13, 2025 · The codebase consists of three primary Python modules that work together to train and evaluate a PPO agent for Super Mario Bros. 0 with CUDA 11. The following diagram illustrates the component structure and their dependencies: Nov 13, 2025 · Purpose and Scope This document provides a complete technical reference for test_model. py and is executed via command-line arguments or the installed script entry point. gxsu, ov8, gxqkb, 7esge, tyiskh, bkjpt, otxuw, zrg, n3ugtqy, vybfbn, \