[CVPR2026]🚀🚀🚀Official code for the paper "YOLO-Master: MOE-Accelerated with Specialized Transformers for Enhanced Real-time Detection." *(YOLO = You Only Look Once)* 🔥🔥🔥

Code for Xiaomi-Robotics-1

🔥 Quo Vadis, World Modeling? Towards Interactive World Proxies for Continually Improving Agents

Code for SM4RT: Learning Structured Motion Geometry for 4D Reconstruction

ONNX-compatible LightGlue: Local Feature Matching at Light Speed. Supports TensorRT, OpenVINO

Official PyTorch implementation of "Let RGB Be the Language of Vision".

BoQ: A Place is Worth a Bag of learnable Queries (CVPR 2024)

Self-supervised learning for spatial perception

Non-invasive decoding of typed sentences from MEG and EEG brain recordings using a convolutional encoder, transformer, and character-level language model.

Official Implementation of "Multi-View Point Tracking as Geometric Supervision for 4D Video Generation"

[ICLR 25, TPAMI 26, CVPR 26] Track-On: Online Point Tracking with Memory

[CVPR'26 Highlight] Official Code for “V²-SAM: Marrying SAM2 with Multi-Prompt Experts for Cross-View Object Correspondence”

[CVPR26] MuM's a pretty good feature extractor for 3D tasks, probably the best.

[NeurIPS'23] Emergent Correspondence from Image Diffusion

Generative 3D Earth Model by AMap-cvlab

[CVPR 2026 Highlight] MV-RoMa: From Pairwise Matching into Multi-View Track Reconstruction

GLUEMAP: Global Structure-from-Motion Meets Feedforward Reconstruction

Official implementation of Déjà View: Looping Transformers for Multi-View 3D Reconstruction

Deep functional map codebase for 3D shape matching with a 33× faster batched solver, multiple method implementations, and evaluation metrics.

TriSplat: Simulation-Ready Feed-Forward 3D Scene Reconstruction

Official implementation of "GARD: Geometry-Aware Representation Denoising for Robust Multi-view 3D Reconstruction"

[CVPR 2026 Oral] VGGT Omega

A curated, continuously updated reading list, paper blogs, and resources for World Action Models (WAMs) in embodied AI.

A theoretical reconstruction of the Claude Mythos architecture, built from first principles using the available research literature.

[CVPR 2026 (Highlight)] Scal3R: Scalable Test-Time Training for Large-Scale 3D Reconstruction

Allen Institute for AI: WildDet3D: Scaling Promptable 3D Detection in the Wild

Mastering Diverse Domains through World Models

pySLAM is a hybrid Python/C++ Visual SLAM pipeline supporting monocular, stereo, and RGB-D cameras. It provides a broad set of modern local and global feature extractors, multiple loop-closure stra…