Anurag Ghosh

dex-lang. Research language for array processing in the Haskell/ML family

spark. :sparkles: An advanced 3D Gaussian Splatting renderer for THREE.js

phyco. Checkpoint and evaluation code for PhyCo [CVPR 2026]

gluemap. GLUEMAP: Global Structure-from-Motion Meets Feedforward Reconstruction

OneVL_training. Python

prix. PRIX: Learning to Plan from Raw Pixels for End-to-End Autonomous Driving

FrameCrafter. [ECCV 2026] FrameCrafter: Novel View Synthesis as Video Completion

street_crafter. [CVPR 2025] StreetCrafter: Street View Synthesis with Controllable Video Diffusion Models

OpenVO. Implementation of Open-World Visual Odometry with Temporal Dynamics Awareness (CVPR'26)

onevl. Python

starVLA. StarVLA: A Lego-like Codebase for Vision-Language-Action Model Developing

vggt-omega. [CVPR 2026 Oral] VGGT Omega

dspy. DSPy: The framework for programming—not prompting—language models

py123d. 123D: A Unified Library for Multi-Modal Autonomous Driving Data

ClickHouse. ClickHouse® is a real-time analytics database management system

colmap. COLMAP - Structure-from-Motion and Multi-View Stereo

WorkZone3D. A repository for the multimodal WorkZone3D dataset

tuna-2. Official implementation of Tuna-2: Pixel Embeddings Beat Vision Encoders for Unified Understanding and Generation

fail2drive. Fail2Drive: Benchmarking Closed-Loop Driving Generalization

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

waymax. A JAX-based simulator for autonomous driving research.

LoMa. [ECCV 2026 Oral] LoMa: Local Feature Matching Revisited

much-ado-about-noising. A optimized PyTorch framework for behavior cloning with flow related generative models.

NuPlanQA.

reachy_mini. Reachy Mini's SDK

MOTIP. [CVPR 2025] Multiple Object Tracking as ID Prediction

OLMo-core. PyTorch building blocks for the OLMo ecosystem

LTX-2. Official Python inference and LoRA trainer package for the LTX-2 audio–video generative model.

sam-audio. The repository provides code for running inference with the Meta Segment Anything Audio Model (SAM-Audio), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

cosmos-transfer2.5. Cosmos-Transfer2.5, built on top of Cosmos-Predict2.5, produces high-quality world simulations conditioned on multiple spatial control inputs.

Cosmos-Drive-Dreams. Cosmos-Drive-Dreams: Scalable Synthetic Driving Data Generation with World Foundation Models

pyvrs. Python interface for https//github.com/facebookresearch/vrs.

attention-map-diffusers. 🚀 Cross attention map tools for huggingface/diffusers

ConceptAttention. ConceptAttention: A method for interpreting multi-modal diffusion transformers.

sam3. The repository provides code for running inference and finetuning with the Meta Segment Anything Model 3 (SAM 3), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

OpenEMMA. OpenEMMA, a permissively licensed open source "reproduction" of Waymo’s EMMA model.

CaRL. [CoRL 2025] CaRL: Learning Scalable Planning Policies with Simple Rewards

InteractiveClosedLoop.

alpasim. AlpaSim is an open-source autonomous vehicle simulation platform designed for development and testing of end-to-end AV policies

mpsfm. MP-SfM: Monocular Surface Priors for Robust Structure-from-Motion (CVPR 2025)

interPlan. Python

ryugraph. Ryu, a fork of Kuzu, is an Embedded Property Graph Database built for speed with vector search and full-text search built in. Implements Cypher.

torchax. torchax is a PyTorch frontend for JAX. It gives JAX the ability to author JAX programs using familiar PyTorch syntax. It also provides JAX-Pytorch interoperability, meaning, one can mix JAX & Pytorch syntax together when authoring ML programs, and run it in every hardware JAX can run.

GEN3C. [CVPR 2025 Highlight] GEN3C: 3D-Informed World-Consistent Video Generation with Precise Camera Control

openpi. Python

Depth-Anything-V2. [NeurIPS 2024] Depth Anything V2. A More Capable Foundation Model for Monocular Depth Estimation

MoGe. [CVPR'25 Oral] MoGe: Unlocking Accurate Monocular Geometry Estimation for Open-Domain Images with Optimal Training Supervision

lqrax. GPU-friendly, auto-differentiable LQR solver with JAX.

xinfer. Blazing-fast LLM inference in pure Rust. No PyTorch and Python runtime.

gpt-oss. gpt-oss-120b and gpt-oss-20b are two open-weight language models by OpenAI

harmony. Renderer for the harmony response format to be used with gpt-oss

planTF. [ICRA'2024] Rethinking Imitation-based Planner for Autonomous Driving

pluto. PLUTO: Push the Limit of Imitation Learning-based Planning for Autonomous Driving

safe-sim. Python

nano-vllm. Nano vLLM

Diffusion-Planner. [ICLR 2025 Oral] The official implementation of "Diffusion-Based Planning for Autonomous Driving with Flexible Guidance"

pyslam. 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 strategies, a volumetric reconstruction module, integrated depth-prediction models, and semantic segmentation capabilities for enhanced scene understanding.

Prior-Depth-Anything. Python

fastmap. A fast and simple structure from motion pipeline written in Pytorch.

Patch_Scaling. Official implementation of Scaling Laws in Patchification: An Image Is Worth 50,176 Tokens And More

aerial-megadepth. [CVPR 2025] AerialMegaDepth: Learning Aerial-Ground Reconstruction and View Synthesis

infinigen. Infinite Photorealistic Worlds using Procedural Generation

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

imm. Official implementation of Inductive Moment Matching

MegaLoc. An image retrieval model for any localization task

gpudrive. 1 million FPS multi-agent driving simulator

samurai. Official repository of "SAMURAI: Adapting Segment Anything Model for Zero-Shot Visual Tracking with Motion-Aware Memory"

mamba. The Fast Cross-Platform Package Manager

UniDepth. Universal Monocular Metric Depth Estimation

Instance-Warp. [WACV 2025] Instance-Level Image Warping for Domain Adaptation

inverse_painting. Inverse Painting: Reconstructing The Painting Process (SIGGRAPH ASIA 2024)

rope-vit. [ECCV 2024] Official PyTorch implementation of RoPE-ViT "Rotary Position Embedding for Vision Transformer"

detrex. detrex is a research platform for DETR-based object detection, segmentation, pose estimation and other visual recognition tasks.

vismatch. Wrapper of 50+ image matching models with a unified interface

ml-sigmoid-attention. Python

droid_metric. run DROID-SLAM with Metric3D to improve monocular performance

drivestudio. A 3DGS framework for omni urban scene reconstruction and simulation.

mesh-data-synthesizer. Uses Unreal Engine & Cesium to generate large synthetic dataset from 3D meshes. Enables machine learning tasks like Visual Place Recognition read more in our paper on this: https://meshvpr.github.io

SCLIP. Official implementation of SCLIP: Rethinking Self-Attention for Dense Vision-Language Inference

glasbey. Algorithmically create or extend categorical colour palettes

rerun. Visualize, query, and stream to train on multimodal robotics data.

Grendel-GS. [ICLR 2025 Oral] On Scaling Up 3D Gaussian Splatting Training

Stable-DINO. [ICCV 2023] Official implementation of the paper "Detection Transformer with Stable Matching"

glomap. [DEPRECATED] GLOMAP - Global Structured-from-Motion Revisited

einx. Universal Notation for Tensor Operations in Python

assetto_corsa_gym. Assetto Corsa OpenAI Gym Environment

x360dataset-kit. Python

Marigold. [CVPR 2024 - Oral, Best Paper Award Candidate] Marigold: Repurposing Diffusion-Based Image Generators for Monocular Depth Estimation

sigma-gpt. σ-GPT: A New Approach to Autoregressive Models

mixed-resolution-vit. Python

roadwork-dataset. Repository that contains simple scripts to use ROADWork dataset.

RPG-DiffusionMaster. [ICML 2024] Mastering Text-to-Image Diffusion: Recaptioning, Planning, and Generating with Multimodal LLMs (RPG)

Context-Cluster. [ICLR 2023 Oral] Image as Set of Points

scaling_on_scales. When do we not need larger vision models?

pylot. Modular autonomous driving platform running on the CARLA simulator and real-world vehicles.

cad. Content-Adaptive Downsampling in Convolutional Neural Networks (CVPR 2023 Workshop on Efficient Deep Learning for Computer Vision)

gaussian_surfels. [SIGGRAPH'24] Implementations for "High-quality Surface Reconstruction using Gaussian Surfels".

DROID-SLAM. Python

MS-DOS. The original sources of MS-DOS 1.25, 2.0, and 4.0 for reference purposes

assignment4. 3D Gaussian Splatting and Diffusion Guided Optimization

smalldiffusion. Simple and readable code for training and sampling from diffusion models

SKS-Homography. Implementation of PAMI 2025 paper "Fast and Interpretable 2D Homography Decomposition: Similarity-Kernel-Similarity and Affine-Core-Affine Transformations"

OpenTrafficCam3D. Source code for paper Toward Planet-Wide Traffic Camera Calibration (WACV 2024)

dust3r. DUSt3R: Geometric 3D Vision Made Easy

vllm-code-harness. Run code inference-only benchmarks quickly using vLLM

assignment3. Volume Rendering, Neural Radiance Fields, Neural Surfaces

doppelgangers. Doppelgangers: Learning to Disambiguate Images of Similar Structures

GPT-Driver. Learning to Drive with GPT

StreamingForecasting. [IROS 2023] "Streaming Motion Forecasting for Autonomous Driving"

2pcnet. Python

OpenSeeD. [ICCV 2023] Official implementation of the paper "A Simple Framework for Open-Vocabulary Segmentation and Detection"

img2dataset. Easily turn large sets of image urls to an image dataset. Can download, resize and package 100M urls in 20h on one machine.

Chanakya. Learning Runtime Decisions for Adaptive Real-Time Perception

voicecamo. Code for the paper Real-Time Neural Voice Camouflage

gaussian-splatting. Original reference implementation of "3D Gaussian Splatting for Real-Time Radiance Field Rendering"

two-plane-prior. Learned Two-Plane Perspective Prior based Image Resampling for Efficient Object Detection

DeepImageBlending. This is a Pytorch implementation of deep image blending

hiera. Hiera: A fast, powerful, and simple hierarchical vision transformer.

lzu. Code for Learning to Zoom and Unzoom (CVPR 2023)

submitit. Python 3.8+ toolbox for submitting jobs to Slurm

SAM-Tool. 利用Segment Anything(SAM)模型进行快速标注

salt. Segment Anything Labelling Tool

pyboreas. Devkit for the Boreas autonomous driving dataset.

HAMS. Code for Automated License Testing from the HAMS group @ Microsoft Research, India

TaskMatrix. Python

llama.cpp. LLM inference in C/C++

slahmr. Python

yolov7. Implementation of paper - YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors

llama. Inference code for Llama models

arxiv-submission-sanitizer-flattener. Simple Python scripts to clean up and flatten ArXiv LaTeX submissions.

theseus. A library for differentiable nonlinear optimization

imagenetx. understanding model mistakes with human annotations

kornia. 🐍 Geometric Computer Vision Library for Spatial AI

VanishingPoint_HoughTransform_GaussianSphere. Official implementation: Deep vanishing point detection: Geometric priors make dataset variations vanish, CVPR'22.

av2-api. Argoverse 2: Next generation datasets for self-driving perception and forecasting.

pytorch3d. PyTorch3D is FAIR's library of reusable components for deep learning with 3D data

Detic. Code release for "Detecting Twenty-thousand Classes using Image-level Supervision".

mmcv. OpenMMLab Computer Vision Foundation

mmtracking. OpenMMLab Video Perception Toolbox. It supports Video Object Detection (VID), Multiple Object Tracking (MOT), Single Object Tracking (SOT), Video Instance Segmentation (VIS) with a unified framework.

on-demand-fork. On-demand-fork

cvxpylayers. Differentiable convex optimization layers

mmdetection. OpenMMLab Detection Toolbox and Benchmark

Verified-Telemetry. Azure Verified Telemetry for IoT is a state-of-the-art solution to seamlessly determine the health of the sensor in real-time.

HTML4Vision. A simple HTML visualization tool for computer vision research :hammer_and_wrench:

ApproxDet. Python

UniverseNet. USB: Universal-Scale Object Detection Benchmark (BMVC 2022)

tide. A General Toolbox for Identifying Object Detection Errors

cocoapi. COCO API - Dataset @ http://cocodataset.org/

jetson_stats. 📊 Simple package for monitoring and control your NVIDIA Jetson [Orin, Xavier, Nano, TX] series

sAP. Code for Towards Streaming Perception (ECCV 2020) :car:

jsondaora. Interoperates dataclasses and TypedDict annotations with json objects for python

jsonschema-typed. Use JSON Schema for type checking in Python

serve. Serve, optimize and scale PyTorch models in production

android_rinex. This repository contains a python script that converts logs from Google's GNSS measurement tools to RINEX

VoTT. Visual Object Tagging Tool: An electron app for building end to end Object Detection Models from Images and Videos.

awesome-computer-vision-in-sports. A comprehensive list of papers on computer vision in sports

pytorch_divcolor. Diverse Colorization in Torch

NeuralNetwork-Viterbi. Python

tcfpn-isba. Weakly-Supervised Action Segmentation with Iterative Soft Boundary Assignment (CVPR 2018)

labelImg. LabelImg is now part of the Label Studio community. The popular image annotation tool created by Tzutalin is no longer actively being developed, but you can check out Label Studio, the open source data labeling tool for images, text, hypertext, audio, video and time-series data.

keras-js. Run Keras models in the browser, with GPU support using WebGL

trytond. Mirror of trytond

netvlad. NetVLAD: CNN architecture for weakly supervised place recognition

image-forensics. Java

cvpr2015. Jupyter Notebook

imagenet-multiGPU.torch. an imagenet example in torch.

Hindi-Latex-Template. TeX

visualindex. A simple demo of visual object matching using VLFeat

CNTK. Microsoft Cognitive Toolkit (CNTK), an open source deep-learning toolkit

EigenLibSVM. A wrapper for LibSVM that lets you train SVM's directly on Eigen library matrices in C++

CanvasInput. HTML5 Canvas Text Input

PureDarwin. Darwin is the Open Source core of macOS, and PureDarwin is a community project to extend Darwin into a complete, usable operating system.

Automated-Essay-Grading. Python

electron. :electron: Build cross-platform desktop apps with JavaScript, HTML, and CSS

rapt. Robots Are People Too, a platformer in HTML5

console2-solarized. The Solarized theme ported to Console2

Cello. Higher level programming in C

todo. My Terminal todo app

gdc. GNU D Compiler