hehao

FastVideo. A unified inference and post-training framework for accelerated video generation.

hermes-agent. The agent that grows with you

demoparser. Counter-Strike 2 replay parser for Python and JavaScript

LongLive-RAG. Official Implementation of LongLive-RAG: A general retrieval-augmented framework for long video generation.

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

SenseNova-U1. SenseNova-U series: Native Unified Paradigm with NEO-unify from the First Principles

cli. The official Lark/Feishu CLI tool, maintained by the larksuite team — built for humans and AI Agents. Covers core business domains including Messenger, Docs, Base, Sheets, Calendar, Mail, Tasks, Meetings, and more, with 200+ commands and 20+ AI Agent Skills.

cosmos-reason2. Cosmos-Reason2 models understand the physical common sense and generate appropriate embodied decisions in natural language through long chain-of-thought reasoning processes.

ABot-PhysWorld. Python

JTA-Dataset. Python

InternUtopia. A simulation platform for versatile Embodied AI research and developments.

UnrealClaude. Claude Code CLI integration for Unreal Engine 5.7 - Get AI coding assistance with built-in UE5.7 documentation context directly in the editor.

WorldScore. Official implementation for WorldScore: A Unified Evaluation Benchmark for World Generation

madpose. [CVPR 2025 Highlight] Official implementation of the solvers and estimators proposed in the paper "Relative Pose Estimation through Affine Corrections of Monocular Depth Priors"

VideoGPA. [ICML'26] VideoGPA is a self-supervised framework that enhances 3D consistency in Video Diffusion Models.

vipe. ViPE: Video Pose Engine for Geometric 3D Perception

lingbot-vla. A Pragmatic VLA Foundation Model

lingbot-depth. Masked Depth Modeling for Spatial Perception

VLMEvalKit. Open-source evaluation toolkit of large multi-modality models (LMMs), support 220+ LMMs, 80+ benchmarks

MMSI-Bench. [ICLR 2026] MMSI-Bench: A Benchmark for Multi-Image Spatial Intelligence

Structured3D. [ECCV'20] Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling

MagiAttention. A Distributed Attention Towards Linear Scalability for Ultra-Long Context, Heterogeneous Data Training

G2VLM. [CVPR 2026] G2VLM: Geometry Grounded Vision Language Model with Unified 3D Reconstruction and Spatial Reasoning

evo. Python package for the evaluation of odometry and SLAM

DOVER. [ICCV 2023, Official Code] for paper "Exploring Video Quality Assessment on User Generated Contents from Aesthetic and Technical Perspectives". Official Weights and Demos provided.

sam-3d-objects. SAM 3D Objects

Depth-Anything-3. Depth Anything 3

MAGI-1. MAGI-1: Autoregressive Video Generation at Scale

VST. [ECCV2026] Visual Spatial Tuning

RollingForcing. [ICLR 2026] Official Repo for Rolling Forcing: Autoregressive Long Video Diffusion in Real Time

LongLive. Long Video Gen Infrastructure

OmniVinci. OmniVinci is an omni-modal LLM for joint understanding of vision, audio, and language.

xfactor-nvs. Public code for XFactor: Introduces the first geometry-free model to achieve true self-supervised / pose-free Novel View Synthesis (NVS) by learning transferable latent camera pose representations.

Emu3.5. Native Multimodal Models are World Learners

prolificdreamer. ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation (NeurIPS 2023 Spotlight)

DMD2. (NeurIPS 2024 Oral 🔥) Improved Distribution Matching Distillation for Fast Image Synthesis

DepthLM_Official. [ICLR 2026 Oral (top 1.2%)] Official implementation of DepthLM

Unify-Post-Training. Towards a Unified View of Large Language Model Post-Training

ddpo-pytorch. DDPO for finetuning diffusion models, implemented in PyTorch with LoRA support

SocialNavSUB. [CoRL 2025] VLM Benchmark for Social Navigation Scene Understanding

Awesome_Think_With_Images. Resources and paper list for "Thinking with Images for LVLMs". This repository accompanies our survey on how LVLMs can leverage visual information for complex reasoning, planning, and generation.

Waver. Industry-level video foundation model for unified Text-to-Video (T2V) and Image-to-Video (I2V) generation.

YUME. The official code of Yume

VGGT-SLAM. VGGT-SLAM: Dense RGB SLAM Optimized on the SL(4) Manifold

NextStep-1. [🚀 ICLR 2026 Oral] NextStep-1: SOTA Autogressive Image Generation with Continuous Tokens. A research project developed by the StepFun’s Multimodal Intelligence team.

leapvo. [CVPR 2024] LEAP-VO: Long-term Effective Any Point Tracking for Visual Odometry

dinov3. Reference PyTorch implementation and models for DINOv3

BLIP3o. Official implementation of BLIP3o-Series

MoBA. MoBA: Mixture of Block Attention for Long-Context LLMs

HunyuanWorld-1.0. Generating Immersive, Explorable, and Interactive 3D Worlds from Words or Pixels with Hunyuan3D World Model

SpaTrackerV2. [ICCV 2025] SpatialTrackerV2: 3D Point Tracking Made Easy

ao. PyTorch native quantization and sparsity for training and inference

waymo-open-dataset. Waymo Open Dataset

CUT3R. Official implementation of Continuous 3D Perception Model with Persistent State

EmbodiedOcc. [ICCV 2025] Embodied 3D Occupancy Prediction for Vision-based Online Scene Understanding

taskonomy. Taskonomy: Disentangling Task Transfer Learning [Best Paper, CVPR2018]

omnidata. A Scalable Pipeline for Making Steerable Multi-Task Mid-Level Vision Datasets from 3D Scans [ICCV 2021]

radial-attention. [NeurIPS 2025] Radial Attention: O(nlogn) Sparse Attention with Energy Decay for Long Video Generation

OpenSfM. Open source Structure-from-Motion pipeline

d2-net. D2-Net: A Trainable CNN for Joint Description and Detection of Local Features

dvgformer. Code for our paper: Learning Camera Movement Control from Real-World Drone Videos

FlashDepth. The official implementation of ICCV'25 paper "FlashDepth: Real-time Streaming Video Depth Estimation at 2K Resolution"

sekai-codebase. [NeurIPS 2025] Sekai: A Video Dataset towards World Exploration

Tar. [NeurIPS 2025] Vision as a Dialect: Unifying Visual Understanding and Generation via Text-Aligned Representations

openpi. Python

OmniGen2. OmniGen2: Exploration to Advanced Multimodal Generation. https://arxiv.org/abs/2506.18871

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

uco3d. Uncommon Objects in 3D dataset

Self-Forcing. Official codebase for "Self Forcing: Bridging Training and Inference in Autoregressive Video Diffusion" (NeurIPS 2025 Spotlight)

kubric. A data generation pipeline for creating semi-realistic synthetic multi-object videos with rich annotations such as instance segmentation masks, depth maps, and optical flow.

SeedVR. Repo for SeedVR2 (ICLR2026) & SeedVR (CVPR2025 Highlight)

WorldMem. [NeurIPS 2025] WorldMem: Long-term Consistent World Simulation with Memory

LAPA. [ICLR 2025] LAPA: Latent Action Pretraining from Videos

UniGeo. UniGeo: Taming Video Diffusion for Unified Consistent Geometry Estimation

RoboMaster. [ICLR’26] Learning Video Generation for Robotic Manipulation with Collaborative Trajectory Control

OpenUni. Python

FlowMo. Official PyTorch implementation of FlowMo.

LoopNav. Python

reverse-engineering-gemma-3n. Reverse Engineering Gemma 3n: Google's New Edge-Optimized Language Model

co-tracker. CoTracker is a model for tracking any point (pixel) on a video.

MMaDA. MMaDA - Open-Sourced Multimodal Large Diffusion Language Models (dLLMs with block diffusion, mixed-CoT, unified RL)

Bagel. Open-source unified multimodal model

FoundationPose. [CVPR 2024 Highlight] FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects

flow_grpo. [NeurIPS 2025] An official implementation of Flow-GRPO: Training Flow Matching Models via Online RL

RealCam-I2V. Python

CamI2V. official repo of paper for "CamI2V: Camera-Controlled Image-to-Video Diffusion Model"

DiffSynth-Studio. Enjoy the magic of Diffusion models!

Voyager. An Open-Ended Embodied Agent with Large Language Models

Magma. [CVPR 2025] Magma: A Foundation Model for Multimodal AI Agents

UI-TARS. Pioneering Automated GUI Interaction with Native Agents

Janus. Janus-Series: Unified Multimodal Understanding and Generation Models

vila-u. [ICLR 2025] VILA-U: a Unified Foundation Model Integrating Visual Understanding and Generation

T2I-R1. [NeurIPS 2025] T2I-R1: Reinforcing Image Generation with Collaborative Semantic-level and Token-level CoT

ml-tarflow. Python

CameraBench. [NeurIPS 2025 Spotlight] Towards Understanding Camera Motions in Any Video

awesome-alignment-of-diffusion-models. [ACM Computing Surveys] The collection of awesome papers on alignment of diffusion models.

REPA-E. [ICCV 2025] Official implementation of the paper: REPA-E: Unlocking VAE for End-to-End Tuning of Latent Diffusion Transformers

MetaSpatial. [ICLR 2026] MetaSpatial leverages reinforcement learning to enhance 3D spatial reasoning in vision-language models (VLMs), enabling more structured, realistic, and adaptive scene generation for applications in the metaverse, AR/VR, and game development.

HunyuanVideo-I2V. HunyuanVideo-I2V: A Customizable Image-to-Video Model based on HunyuanVideo

FramePack. Lets make video diffusion practical!

perception_models. State-of-the-art Image & Video CLIP, Multimodal Large Language Models, and More!

SeedVR. [CVPR2025 Highlight] SeedVR: Seeding Infinity in Diffusion Transformer Towards Generic Video Restoration

ml-matrix3d. [CVPR 2025 Highlight] Matrix3D: Large Photogrammetry Model All-in-One

Geo4D. [ICCV 2025 Highlight] Geo4D: Leveraging Video Generators for Geometric 4D Scene Reconstruction

metamorph. Code for MetaMorph Multimodal Understanding and Generation via Instruction Tuning

ARD. [CVPR 2025 Oral] PyTorch re-implementation for Autoregressive Distillation of Diffusion Transformers (ARD).

SuperPoint. Efficient neural feature detector and descriptor

GigaTok. [ICCV 2025] Official repo for "GigaTok: Scaling Visual Tokenizers to 3 Billion Parameters for Autoregressive Image Generation"

finetrainers. Scalable and memory-optimized training of diffusion models

DDT. [CVPR 2026] DDT: Decoupled Diffusion Transformer

LLaMA-Mesh. Unifying 3D Mesh Generation with Language Models

PixelFlow. Pixel-Space Generative Models

deep-rl-class. This repo contains the Hugging Face Deep Reinforcement Learning Course.

unimatch. [TPAMI'23] Unifying Flow, Stereo and Depth Estimation

muggled_dpt. Muggled DPT: Depth estimation without the magic

croco. Python

PAR. [CVPR2025 Highlight] PAR: Parallelized Autoregressive Visual Generation. https://yuqingwang1029.github.io/PAR-project

GoT. Official repository of "GoT: Unleashing Reasoning Capability of Multimodal Large Language Model for Visual Generation and Editing"

AnimeGamer. [ICCV 2025] AnimeGamer: Infinite Anime Life Simulation with Next Game State Prediction

anycam. Official repository for "AnyCam: Learning to Recover Camera Poses and Intrinsics from Casual Videos" (CVPR 2025)

DAPO. An Open-source RL System from ByteDance Seed and Tsinghua AIR

SegAnyMo. [CVPR 2025] Code for Segment Any Motion in Videos

dso. [ICCV 2025] DSO: Aligning 3D Generators with Simulation Feedback for Physical Soundness

DeepSeek-Math. DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models

FAR. Code for: "Long-Context Autoregressive Video Modeling with Next-Frame Prediction"

Qwen2.5-Omni. Qwen2.5-Omni is an end-to-end multimodal model by Qwen team at Alibaba Cloud, capable of understanding text, audio, vision, video, and performing real-time speech generation.

OpenRLHF. An Easy-to-use, Scalable and High-performance Agentic RL Framework based on Ray (PPO & DAPO & REINFORCE++ & VLM & TIS & vLLM & Ray & Async RL)

carve3d. Code for Carve3D: Improving Multi-view Reconstruction Consistency for Diffusion Models with RL Finetuning

fractalgen. PyTorch implementation of FractalGen https://arxiv.org/abs/2502.17437

RDLM. Official Code Repository for the paper "Continuous Diffusion Model for Language Modeling" (NeurIPS 2025).

UI-TARS-desktop. The Open-Source Multimodal AI Agent Stack: Connecting Cutting-Edge AI Models and Agent Infra

bd3lms. [ICLR 2025 Oral] Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models

LLaDA. Official PyTorch implementation for "Large Language Diffusion Models"

FLARE. Python

Video-T1. [ICCV 2025] Video-T1: Test-Time Scaling for Video Generation

cosmos-reason1. Cosmos-Reason1 models understand the physical common sense and generate appropriate embodied decisions in natural language through long chain-of-thought reasoning processes.

stable-virtual-camera. Stable Virtual Camera: Generative View Synthesis with Diffusion Models

ReCamMaster. [ICCV'25 Best Paper Finalist] ReCamMaster: Camera-Controlled Generative Rendering from A Single Video

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

Step-Video-T2V. Python

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

UniTok. [NeurIPS 2025 Spotlight] A Unified Tokenizer for Visual Generation and Understanding

EasyR1. EasyR1: An Efficient, Scalable, Multi-Modality RL Training Framework based on veRL

Wan2.1. Wan: Open and Advanced Large-Scale Video Generative Models

scaling-with-vocab. [NeurIPS-2024] 📈 Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies https://arxiv.org/abs/2407.13623

nabla-gfn. Official Implementation of Nabla-GFlowNet (ICLR 2025)

open-infra-index. Production-tested AI infrastructure tools for efficient AGI development and community-driven innovation

wonderland. Python

WonderJourney. Python

DiffusionAsShader. [SIGGRAPH 2025] Diffusion as Shader: 3D-aware Video Diffusion for Versatile Video Generation Control

SkyReels-V1. SkyReels V1: The first and most advanced open-source human-centric video foundation model

GS-DiT. Source code for paper GS-DiT: Advancing Video Generation with Pseudo 4D Gaussian Fields through Efficient Dense 3D Point Tracking

WonderWorld. Code release for https://kovenyu.com/WonderWorld/

SynCamMaster. [ICLR'25] SynCamMaster: Synchronizing Multi-Camera Video Generation from Diverse Viewpoints

diamond. DIAMOND (DIffusion As a Model Of eNvironment Dreams) is a reinforcement learning agent trained in a diffusion world model. NeurIPS 2024 Spotlight.

ac3d. AC3D: Analyzing and Improving 3D Camera Control in Video Diffusion Transformers

Lumina-Video. Python

VideoAlign. [NeurIPS 2025] Improving Video Generation with Human Feedback

DB. A PyTorch implementation of "Real-time Scene Text Detection with Differentiable Binarization".

tokenizers. 💥 Fast State-of-the-Art Tokenizers optimized for Research and Production

open-r1. Fully open reproduction of DeepSeek-R1

1Prompt1Story. 🔥ICLR 2025 (Spotlight) One-Prompt-One-Story: Free-Lunch Consistent Text-to-Image Generation Using a Single Prompt

TokenFlow. [CVPR 2025] 🔥 Official impl. of "TokenFlow: Unified Image Tokenizer for Multimodal Understanding and Generation".

Vchitect-2.0. Vchitect-2.0: Parallel Transformer for Scaling Up Video Diffusion Models

Min-SNR-Diffusion-Training. [ICCV 2023] Efficient Diffusion Training via Min-SNR Weighting Strategy

GameFactory. [ICCV 2025] GameFactory: Creating New Games with Generative Interactive Videos

SONAR. SONAR, a new multilingual and multimodal fixed-size sentence embedding space, with a full suite of speech and text encoders and decoders.

LlamaV-o1. [ACL 2025 🔥] Rethinking Step-by-step Visual Reasoning in LLMs

large_concept_model. Large Concept Models: Language modeling in a sentence representation space

MegaSynth. Code for MegaSynth: Scaling Up 3D Scene Reconstruction with Synthesized Data (CVPR 2025)

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

cosmos. NVIDIA Cosmos is an open platform of world models, datasets, and tools that enables developers to build Physical AI for robots, autonomous vehicles, smart infrastructure, and more.

1xgpt. world modeling challenge for humanoid robots

Sa2VA. Official Repo For Pixel-LLM Codebase: Sa2VA (PAMI-26), SAMTok (CVPR-26), VRT (Arxiv-25), SaSaSa2VA (1-st solution for LSVOS)

VisionReward. [AAAI 2026] VisionReward: Fine-Grained Multi-Dimensional Human Preference Learning for Image and Video Generation

REPA. [ICLR'25 Oral] Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think

LightningDiT. [CVPR 2025 Oral] Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Models

causalfusion. Python

sparse_coding. Using sparse coding to find distributed representations used by neural networks.

ml-depth-pro. Depth Pro: Sharp Monocular Metric Depth in Less Than a Second.

UniDepth. Universal Monocular Metric Depth Estimation

ZoeDepth. Metric depth estimation from a single image

Metric3D. The repo for "Metric3D: Towards Zero-shot Metric 3D Prediction from A Single Image" and "Metric3Dv2: A Versatile Monocular Geometric Foundation Model..."

DeepSeek-V3. Python

PromptDA. [CVPR 2025] Prompt Depth Anything

VideoVAEPlus. [ICCV 2025] VideoVAE+: Large Motion Video Autoencoding with Cross-modal Video VAE

open-muse. Open reproduction of MUSE for fast text2image generation.

DROID-Splat. End-to-End SLAM with camera calibration, monocular prior integration and dense Rendering

VideoDPO. Official Implementation of VideoDPO

VidTok. a family of versatile and state-of-the-art video tokenizers.

NOVA. [ICLR 2025] Autoregressive Video Generation without Vector Quantization

MVAR. Python

HumanVid. [NeurIPS D&B Track 2024] Official implementation of HumanVid

DeepSeek-VL2. DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding

StreamDiffusion. StreamDiffusion: A Pipeline-Level Solution for Real-Time Interactive Generation

genex. Generative World Explorer

Owl.

streamv2v. Official Pytorch implementation of StreamV2V.

distillnerf. [NeurIPS 2024] DistillNeRF: Perceiving 3D Scenes from Single-Glance Images by Distilling Neural Fields and Foundation Model Features

flow_matching. A PyTorch library for implementing flow matching algorithms, featuring continuous and discrete flow matching implementations. It includes practical examples for both text and image modalities.