datawhalechina / every-embodied
仅需Python基础，从0构建自己的具身智能机器人；从0逐步构建VLA/OpenVLA/SmolVLA/Pi0， 深入理解具身智能
datawhalechina / every-embodied
JethroJames / awesome-robots-icl
Companion repository for “In-Context Learning for Robots: Methods and Applications” — methods, datasets, and benchmarks organized by taxonomy and release date.
kyutai-labs / reki
REKI: a 2.4B vision-language-action model pretrained on public data only, with a memory of its recent past at negligible cost.
starVLA / starVLA
StarVLA: A Lego-like Codebase for Vision-Language-Action Model Developing
robocurve / inspect-robots
Open source evals for physical AI. Run any LLM/VLA on any arm/humanoid against any real/sim benchmark.
dexmal / opendm
An Open-World Foundation Model for General-Purpose Embodied Intelligence.
FluxVLA / FluxVLA
An all-in-one VLA engineering platform for embodied AI — from data to real-robot deployment.
xiaomi-research / recogdrive
[ICLR 2026] ReCogDrive: A Reinforced Cognitive Framework for End-to-End Autonomous Driving
starVLA / VLAct
[NeurIPS 2026] Beyond Data Scaling: Representation-Centric Continued Pre-training for Vision-Language-Action Models
DexForce / UNITAS
Official codebase for Unitas: A 3D-Native World Action Model for Embodied Manipulation
DAGroup-PKU / PyRUA-Lean
Fewer Tokens, Better Action: GPT-6 Astra Robot Agents with 14% Higher Success but 65% Fewer Tokens
thu-ml / Motus
Official code of Motus: A Unified Latent Action World Model
H-EmbodVis / TurboVLA
TurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM
ginwind / VLA-JEPA
[ECCV 2026] VLA-JEPA: Enhancing Vision-Language-Action Model with Latent World Model
OpenMOSS / EasyWAM
A unified framework for training, fine-tuning, and evaluating World Action Models
hku-sail / StreamPI
[NeurIPS 2026] Official code of "StreamPI: Streaming Multimodal Temporal Modeling for Vision-Language-Action Models"
OpenBMB / SimpleNav
SimpleNav: A simple, unified, reproducible, and extensible framework for navigation VLA research.
rzc156 / dreamforge
Turn world-model imagination into corrective training data for robot policies — a single-GPU, falsifiable pipeline (SmolVLA · LIBERO · CEM · paired simulator replays).
dk2472780158-ctrl / piper-dual-arm-smolvla-towel-folding
SmolVLA (SmolVLM2-500M) vision-language-action on real dual AgileX Piper arms — continuous authority blending on a proven ACT safety base folds a real towel. One command runs the full 750-step skill (status=0). Full illustrated course included
OpenBMB / SimpleMemVLA
Native-video memory for vision-language-action models, using timestamped visual history and exact streaming inference for long-horizon robot manipulation.
scy-v / VLA-Precision
VLA-Precision: Asymmetric Co-Bootstrapping for Efficient Real-World Online RL of Vision-Language-Action Models
fendouai / awesome-embodied-ai
2,000 curated embodied AI and robotics resources with bilingual metadata and transparent ranking
3587jjh / HuRo
HuRo: Robotizing Human Videos for Scalable VLA Pretraining (CoRL 2026)
GehaoZhang6 / Grounded-Action-Model
Grounded Action Model: 3D Grounding as a Foundation for Robotics
marmotlab / RL2-VLA
RL²-VLA: Adaptive RL Latent Compositional Steering with Test-Time Scaling for Vision-Language-Action Models
haha-yuki-haha / AUGUstus
🚗 AUGUstus: Autonomous Driving VLA Model with Unified Generation and Understanding. 🏋️ 512 H800 GPUs · ~2 days · 318K+ video hours · 513B+ training tokens. ⚡ 12 Hz on-vehicle deployment. SFT + GRPO + NAVSIM.
Grinffin / PhaseLoRA
Official implementation of PhaseLoRA: control-regime-conditioned low-rank adaptation for continuous-action vision-language-action policies, built on OpenPI.
cair-vinuni / FoldQuantVLA
FoldQuant: offline folding for W4A4 / W8A8 vision-language-action inference on edge GPUs
AureleoPKU / DeCAL
Official Implementation for “DeCAL: Towards Physically-Grounded Dexterous Vision-Language-Action Models via Contact-Aware Latent Co-Imagination” (CoRL 2026)