Awesome AI Agents › Generative Agents Simulation
Odyssey is an advanced framework designed to empower Minecraft agents with open-world skills, enabling them to explore and interact with the vast Minecraft environment more effectively. Unlike previous efforts that focused on basic programmatic tasks such as material collection and tool crafting, Odyssey introduces a comprehensive approach that leverages a large language model (LLM) to enhance agent capabilities. The framework is built upon the Voyager framework and includes three main components: an interactive agent equipped with a skill library containing 40 primitive skills and 183 compositional skills, a fine-tuned LLaMA-3 model trained on a large dataset of over 390,000 instruction entries derived from the Minecraft Wiki, and a new benchmark designed to evaluate agent capabilities across long-term planning, dynamic-immediate planning, and autonomous exploration tasks. Odyssey's skill library allows agents to perform complex sequences of actions, facilitating the discovery of diverse gameplay opportunities beyond traditional goals like obtaining diamonds. The fine-tuned LLaMA-3 model enhances the agent's understanding and decision-making by providing rich contextual knowledge from the Minecraft Wiki. The benchmark tasks assess the agents' ability to plan and act autonomously in an open-world setting, pushing the boundaries of what AI agents can achieve in Minecraft. The project also includes tools for crawling Minecraft Wiki data, a comprehensive skill library for automated item collection and crafting, and a multi-agent framework for parallelized planning and acting. The codebase supports deployment on various operating systems and provides detailed installation instructions for Python, Node.js, and Minecraft server setup. Demonstration videos showcase the agents performing tasks such as mining diamonds, crafting weapons, combat, farming, and autonomous exploration. All datasets, model weights, and code are publicly available to encourage further research and development in autonomous agent solutions for open-world environments.
https://github.com/zju-vipa/Odyssey
This project provides a simulation framework for generative agents that mimic believable human behaviors within an interactive game environment, enabling running, saving, replaying, and demonstrating agent-based simulations.
Voyager is an open-ended embodied agent powered by large language models that autonomously explores and learns diverse skills in Minecraft through lifelong learning and an evolving skill library.
SFighterAI is a deep reinforcement learning-based AI agent designed to master and beat the final boss in Street Fighter II: Special Champion Edition using only game screen pixel data.
A comprehensive Godot framework for creating smooth and complex AI motion using steering behaviors in both 2D and 3D games.
WorldX turns a single sentence into a running simulated world, generating maps and characters with LLMs and letting autonomous agents act, remember and form stories inside it.
Control room for LLM agent societies: pause and scrub a running simulation, question individual residents, inject instructions into the next step, and export replayable experiment packs.
PyGame Learning Environment (PLE) is a Python-based reinforcement learning platform that provides a variety of games and tools to facilitate the development and testing of reinforcement learning algorithms.
A locally runnable, low-cost implementation of generative agents simulating human-like behavior in an interactive Dungeons & Dragons setting using large language models.