The Official Implementation of "ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow"

AI agents running research on single-GPU nanochat training automatically

RLinf: Reinforcement Learning Infrastructure for Embodied and Agentic AI

[ICLR 2023] FluidLab: A Differentiable Environment for Benchmarking Complex Fluid Manipulation

RoboVerse: Towards a Unified Platform, Dataset and Benchmark for Scalable and Generalizable Robot Learning

[ICML 2025] The Official Implementation of "Efficient Robotic Policy Learning via Latent Space Backward Planning"

[ICLR 2025] The offical implementation of "PSEC: Skill Expansion and Composition in Parameter Space", a new framework designed to facilitate efficient and flexible skill expansion and composition, …

[ICRA 2025] H2O+: An Improved Framework for Hybrid Offline-and-Online RL with Dynamics Gaps.

[ICLR 2025 Oral] The official implementation of "Diffusion-Based Planning for Autonomous Driving with Flexible Guidance"

[ICRA'25] H2O+: An Improved Framework for Hybrid Offline-and-Online RL with Dynamics Gaps

[CVPR 2025] The offical Implementation of "Universal Actions for Enhanced Embodied Foundation Models"

CARLA Sim2Real Environment

A PyTorch library for implementing flow matching algorithms, featuring continuous and discrete flow matching implementations. It includes practical examples for both text and image modalities.

Official pytorch implementation of the paper: "An Edit Friendly DDPM Noise Space: Inversion and Manipulations". CVPR 2024.

[RSS 2022 Best Paper, Best Student Paper Finalist] Iterative Residual Policy for Goal-Conditioned Dynamic Manipulation of Deformable Objects

[AAMAS'26] xTED: Cross-Domain Adaptation via Diffusion-Based Trajectory Editing

[AAAI'25 Oral] Are Expressive Models Truly Necessary for Offline RL?

Manipulation Skill Framework, an open source GPU parallelized robotics simulator and benchmark

[ICML 2024] The offical implementation of A2PR, a simple way to achieve SOTA in offline reinforcement learning with an adaptive advantage-guided policy regularization method, in Pytorch

[ICRA 2024] Official codes for paper: Continual Driving Policy Optimization with Closed-Loop Individualized Curricula.

[ICML 2024] The offical Implementation of "DecisionNCE: Embodied Multimodal Representations via Implicit Preference Learning"

[ICLR 2024] The official implementation of "Safe Offline Reinforcement Learning with Feasibility-Guided Diffusion Model"

Octo is a transformer-based robot policy trained on a diverse mix of 800k robot trajectories.

My personal homepage

[IJCAI'24 Oral] An index of algorithms, approaches, and systems on cross-domain policy transfer for embodied agents