Models

Research

Official implementation of CoPa: General Robotic Manipulation through Spatial Constraints of Parts with Foundation Models

Research

A cross-embodiment policy framework for robot learning.

Research

DEAS Isaac GR00T provides setup resources combining DEAS, Isaac GR00T and RoboCasa.

Open source

The offical Implementation of "DecisionNCE: Embodied Multimodal Representations via Implicit Preference Learning"

Research

Deep RL Grasping provides training and evaluation code for depth/RGB-D robotic grasping policies.

Research

Dexora provides VLA policy resources for high-degree-of-freedom bimanual dexterity.

Research

Official code and data from DexWM ("World Models Can Leverage Human Videos for Dexterous Manipulation").

Research

DreamX-Phi provides an action-conditioned video world model for robotic manipulation.

Research

DynamicVLA provides research resources for dynamic-object robotic manipulation.

Research

EAI2025 Project combines 3D semantic mapping and VLA deployment for manipulation.

Open source

The first framework enabling humanoid robots to learn whole-body loco-manipulation from egocentric human demos

Research

Ego-World JEPA studies ego/world latent factorization for planning on PushT.

Research

Emma-X: An Embodied Multimodal Action Model with Grounded Chain of Thought and Look-ahead Spatial Reasoning

Research

EquiDexFlow provides contact-grounded generative flows for dexterous grasping.

Research

Official Implementation of "EVA: Aligning Video World Models with Executable Robot Actions via Inverse Dynamics Rewards"

Research

EVEWorld studies physical-evolution supervision and temporal consistency in embodied world models.

Research

FLUX 3 Action provides world-action modeling and action-prediction resources for robots and simulators.

Open source

G3Flow: Generative 3D Semantic Flow for Pose-aware and Generalizable Object Manipulation

Research

Towards Scalable Gaussian World Models for Robotic Manipulation

Research

Google DeepMind robotics models connecting perception and action.

Research

Official code of GEM: Generative Supervision Helps Embodied Intelligence

Research

GenRL: Multimodal-foundation world models enable grounding language and video prompts into embodied domains, by turning them into sequences of latent world model states. Latent state sequences can…

Open source

PyTorch Code for "Generalization in Dexterous Manipulation via Geometry-Aware Multi-Task Learning"

Research

A Roadmap to Build World Models for Robot Policy Evaluation