Official implementation of UnifiedReward & [NeurIPS 2025] UnifiedReward-Think & UnifiedReward-Flex

Gym Interface Wrapper for Simulink Models

Explore the Multimodal “Aha Moment” on 2B Model

Suite of PyBullet reinforcement learning environments targeted towards using tactile data as the main form of observation.

Muon is Scalable for LLM Training

Header-Only C++ Library for Graph Representation and Algorithms

"Sequential Dexterity: Chaining Dexterous Policies for Long-Horizon Manipulation" code repository

A generative and self-guided robotic agent that endlessly propose and master new skills.

Lightning-UQ-Box: Uncertainty Quantification for Neural Networks with PyTorch and Lightning

Inference-time scaling of diffusion-based image and video generation models.

[IROS 2025 Best Paper Award Finalist & IEEE TRO 2026] The Large-scale Manipulation Platform for Scalable and Intelligent Embodied Systems

Official implementation of SwiftSketch

robosuite: A Modular Simulation Framework and Benchmark for Robot Learning

[ICLR 2025] Diffusion Feedback Helps CLIP See Better

A contact solver for physics-based simulations involving 👚 shells, 🪵 solids, 🪢 rods, 🧱 rigid bodies and ⏳ sand.

Witness the aha moment of VLM with less than $3.

[CVPR 2025]Lift3D Foundation Policy: Lifting 2D Large-Scale Pretrained Models for Robust 3D Robotic Manipulation

Explorations into whether a transformer with RL can direct a genetic algorithm to converge faster

"4DGen: Grounded 4D Content Generation with Spatial-temporal Consistency", Yuyang Yin*, Dejia Xu*, Zhangyang Wang, Yao Zhao, Yunchao Wei

Official implementation of Diffusion Policy Policy Optimization, arxiv 2024

Minimal reproduction of DeepSeek R1-Zero

Python framework for integrating geospatial data, 3D city modeling, and urban simulation.

[NeurIPS 2025 Spotlight] Reasoning Environments for Reinforcement Learning with Verifiable Rewards

Assistive Gym, a physics-based simulation framework for physical human-robot interaction and robotic assistance.

Simple RL training for reasoning

Janus-Series: Unified Multimodal Understanding and Generation Models

RAGEN leverages reinforcement learning to train LLM reasoning agents in interactive, stochastic environments.

[ICLR 2025] Official implementation of "DiffSplat: Repurposing Image Diffusion Models for Scalable 3D Gaussian Splat Generation".