Stars
Unified World Model Inference & Evaluation Infrastructure
Train a Unitree G1 whole-body motion tracker on a MacBook. World-model RL with MuJoCo on the CPU and MLX on the GPU.
Batched collision-aware IK and trajectory optimization on any GPU (wgpu: Vulkan/Metal/DX12), with a matching CPU backend
Fine-tune LLMs on your Mac with Apple Silicon. SFT, DPO, GRPO, Vision, TTS, STT, Embedding, and OCR fine-tuning — natively on MLX. Unsloth-compatible API.
A CLI for interacting with Google Colab runtimes
WebGPU port of Newton's SolverSteel colored articulated contact path
A lightweight, scaleable rigid-body physics solver
TRELLIS.2 image-to-3D generation, native on Apple Silicon (PyTorch MPS + Metal)
Multi-Joint dynamics with Contact. A general purpose physics simulator.
PolicyIK: High-Precision Inverse Kinematics for Robot Arms with Reinforcement Learning
A toolbox of film styles made in code for Claude Code: 99 ready to use, and your own. Each style is a style prompt plus a demo film made entirely in code by Claude Opus 5.5, with no video model. By…
Hierarchical Franka tabletop pick-and-place for Isaac Lab.
WeVisOne-VL is a powerful VLM by the WeChat Vision Team at Tencent, unifying object-level and pixel-level perception from points and boxes to masks and alpha mattes.
[Assembly NIST] Assembly policy learning for a Franka arm in MuJoCo. Independent implementation of AlphaNIST with PPO, training code, and pretrained policies.
Repository for Timestral, the time-series foundation model by Mistral AI
Reverse engineer anything with agents, from app behavior down to native binaries.
[NeurIPS 2025] SpatialLM: Training Large Language Models for Structured Indoor Modeling
Physical AI skills for coding agents, with field-tested robotics guidance, the robium-ai CLI, and a learning engine.
A Benchmark for Retrieving Papers that Inspire New Research
Implementation of Mimic-Video, Video-Action Models for SOTA Generalizable Robot Control Beyond VLAs
Official PyTorch implementation of VisionHOPE: Visual Backbones as Self-Modifying Learning Systems.
SuperDex brings together a purpose-built physics engine, robotics authoring tools, and a scalable reinforcement learning interface in a unified simulation platform, with VR-based teleoperation and …
Argus is an open source dataset-and-format-agnostic harness for the densest annotation of robot learning data.
Implementation of the recurrent looped transformer proposed by Yifan Zhang of Princeton