[ICCV2023] 🧊FrozenRecon: Pose-free 3D Scene Reconstruction with Frozen Depth Models

PyTorch Extension Library of Optimized Graph Cluster Algorithms

The best ChatGPT that $100 can buy.

Running large language models on a single GPU for throughput-oriented scenarios.

[NeurIPS'23] H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models.

FasterKv.Cache is an in-process hybrid cache library (memory+disk) based on the Microsoft FasterKv.

Awesome-LLM-KV-Cache: A curated list of 📙Awesome LLM KV Cache Papers with Codes.

This repository serves as a comprehensive survey of LLM development, featuring numerous research papers along with their corresponding code links.

Nano vLLM

A curated list of awesome Scriptable scripts and widgets.

Helpful tools and examples for working with flex-attention

Development repository for the Triton language and compiler

Code of WinT3R: Window-Based Streaming Rrconstruction With Camera Token Pool

Fast and memory-efficient exact attention

An ultimately comprehensive paper list of Vision Transformer/Attention, including papers, codes, and related websites

[CVPR 2025 Highlight] Video Depth Anything: Consistent Depth Estimation for Super-Long Videos

Scalable long-context LLM decoding that leverages sparsity—by treating the KV cache as a vector storage system.

[ICML 2024] Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference

[CVPR 2025] MASt3R-SLAM: Real-Time Dense SLAM with 3D Reconstruction Priors

[CVPR 2024] Depth Anything: Unleashing the Power of Large-Scale Unlabeled Data. Foundation Model for Monocular Depth Estimation

[3DV'25] 3D Reconstruction with Spatial Memory

Code for FastVGGT: Training-Free Acceleration of Visual Geometry Transformer

Official Implementation of paper "MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion"

Official implementation of Continuous 3D Perception Model with Persistent State

Dynamic 3D Foundation Model using Causal Transformer

VGGSfM: Visual Geometry Grounded Deep Structure From Motion

[CVPR 2025] LumiNet: Latent Intrinsics Meets Diffusion Models for Indoor Scene Relighting

[CVPR 2024 Highlight] SceneTex: High-Quality Texture Synthesis for Indoor Scenes via Diffusion Priors

Google Research