A compact implementation of SGLang, designed to demystify the complexities of modern LLM serving systems.

Sharp Monocular View Synthesis in Less Than a Second

HY-World 1.5: A Systematic Framework for Interactive World Modeling with Real-Time Latency and Geometric Consistency

open-source 3D scanning and processing pipeline

replicAnt - generating annotated images of animals in complex environments with Unreal Engine

(CVPR 2025) From Slow Bidirectional to Fast Autoregressive Video Diffusion Models

A PyTorch native platform for training generative AI models

Ongoing research training transformer models at scale

The simplest, fastest repository for training/finetuning medium-sized GPTs.

Tilus is a tile-level kernel programming language with explicit control over shared memory and registers.

Reference PyTorch implementation and models for DINOv3

ViPE: Video Pose Engine for Geometric 3D Perception

Make your wildest 3D ConvNet dream architectures come true

[CVPR 2025 Best Paper Award] VGGT: Visual Geometry Grounded Transformer

High-Resolution 3D Assets Generation with Large Scale Hunyuan3D Diffusion Models.

Fully open reproduction of DeepSeek-R1

A generative world for general-purpose robotics & embodied AI learning.

CUDA accelerated rasterization of gaussian splatting

Differentiable Rendering Toolkit

[CVPR 2025 - Highlight] Original implementation of "3D Convex Splatting: Radiance Field Rendering with 3D Smooth Convexes"

Release repo for our SLAM Handbook

3D Reconstruction for all

A 3D Gaussian Splatting framework with various derived algorithms and an interactive web viewer

A Python package for calling Slang modules from PyTorch.

Vision3D: A 3D Vision Library built with PyTorch

Wrapper of 37+ image matching models with a unified interface

A paper list of some recent Transformer-based CV works.