world-model

#计算机科学#Collect some World Models for Autonomous Driving (and Robotic) papers.

[NeurIPS 2024] A Generalizable World Model for Autonomous Driving

Code for "TD-MPC2: Scalable, Robust World Models for Continuous Control"

[CVPR 2024 Highlight] Visual Point Cloud Forecasting

A skill-based platform for ROS v.2 with knowledge representating, planning and reasoning

Code for "Hierarchical World Models as Visual Whole-Body Humanoid Controllers"

Build, evaluate and run General Multi-Agent Assistance with ease

Official repository for "iVideoGPT: Interactive VideoGPTs are Scalable World Models" (NeurIPS 2024), https://arxiv.org/abs/2405.15223

#自然语言处理#[NeurIPS 2024] Agent Planning with World Knowledge Model

#大语言模型#A general AI agent framework that can be adapted to various tasks and environments.

#大语言模型#A general AI agent framework that can be adapted to various tasks and environments.

#计算机科学#A comprehensive list of papers for the definition of World Models and using World Models for General Video Generation, Embodied AI, and Autonomous Driving, including papers, codes, and related website...

[ICLR 2025 Oral] PyTorch code for the paper "Open-World Reinforcement Learning over Long Short-Term Imagination"

Official Implementation of DINO-Foresight: Looking into the Future with DINO

#计算机科学#Official Implementation for the paper "SR-AIF: Solving Sparse-Reward Robotic Tasks from Pixels with Active Inference and World Models"

The PyTorch implementation of paper: "AdaWorld: Learning Adaptable World Models with Latent Actions".

Official implementation of Neural Lithography (SIGGRAPH Asia 2023)

A framework for training world models with virtual environments, complete with annotated environment dataset (RetroAct), exploration agent (AutoExplore Agent), and GenieRedux-G - an implementation of ...

Official repo of the ICLR 2025 paper "MMWorld: Towards Multi-discipline Multi-faceted World Model Evaluation in Videos"

#计算机科学#Official code for the ICLR 2025 paper, "Scaling Offline Model-Based RL via Jointly-Optimized World-Action Model Pretraining"