Liangwei Jiang

accio-cu. Hybrid computer use for AI agents on macOS.

RDM. Python

S1-Omni-Image. A Unified Multimodal Model for Scientific Image Understanding and Generation

rich. Rich is a Python library for rich text and beautiful formatting in the terminal.

DeepPaperNote. DeepPaperNote is an agent skill for deep-reading a single paper and generating high-quality Obsidian-style research notes. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.

dailypaper-skills. 用Agent skills打造我的论文流水线

DanceOPD. 🔥 DanceOPD: On-Policy Generative Field Distillation

DeepSpeed. DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.

PiD. PiD: Fast and High-Resolution Latent Decoding with Pixel Diffusion

SeedVR2. [ICLR2026] SeedVR2: One-Step Video Restoration via Diffusion Adversarial Post-Training

HYPIR. Official implementation of HYPIR: Harnessing Diffusion-Yielded Score Priors for Image Restoration (SIGGRAPH 2025)

MARBLE. Multi-Aspect Reward Balance for Diffusion RL

Boogu-Image. Boogu-Image-0.1 is an Apache-2.0 open-source image generation and editing model family that delivers near-closed-source performance with an order of magnitude less data.

i1. Code release for "i1: A Simple and Fully Open Recipe for Strong Text-to-Image Models"

UniRL. UniRL is a Framework for Unified Multimodal Model Reinforcement Learning

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.

D-OPSD. Official Repo of "D-OPSD: On-Policy Self-Distillation for Continuously Tuning Step-Distilled Diffusion Models"

FaceID-6M. Python

OPD. Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe

OPSD. Python

Flow-OPD. Official Repo of "Flow-OPD: On-Policy Distillation for Flow Matching Models"

VAREdit. Python

HiDream-O1-Image. Python

PRISM. Beyond SFT-to-RL: Pre-alignment via Black-BoxOn-Policy Distillation for Multimodal RL

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

SenseNova-SI. [CVPR 2026] Scaling Spatial Intelligence with Multimodal Foundation Models

Flow-Factory. A unified framework for easy reinforcement learning in Flow-Matching models

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

wool_scripts. 收集一些Loon、Surge、QuantumultX、ShadowRocket、Egern的配置与去广告规则。

ProxyResource. 可莉的Loon资源库 | 插件 | 脚本 | 规则

ScaleEdit-12M. ScaleEdit-12M is the largest open-source image editing dataset to date, spanning 23 task families across diverse real and synthetic domains.

PixelSmile. PixelSmile: Fine-grained facial expression editing with continuous control, reduced semantic entanglement, and strong identity preservation.

FIRM-Reward. Trust Your Critic: Robust Reward Modeling and Reinforcement Learning for Faithful Image Editing and Generation

SpatialEdit. [Official Repo] SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing

Logics-Parsing. Python

JoyAI-Image. JoyAI-Image is the unified multimodal foundation model for image understanding, text-to-image generation, and instruction-guided image editing.

OneReward. Python

GEditBench_v2. GEditBench v2: A Human-Aligned Benchmark for General Image Editing

isc2021. Code for the Image similarity challenge.

DeQA-Score. [CVPR 2025] Teaching Large Language Models to Regress Accurate Image Quality Scores using Score Distribution

Q-Align. ③[ICML2024] [IQA, IAA, VQA] All-in-one Foundation Model for visual scoring. Can efficiently fine-tune to downstream datasets.

Self-Flow. [ICML'26] Code and website for Self-Flow: Self-Supervised Flow Matching for Scalable Multi-Modal Synthesis

Q-Insight. Q-Insight Family: Q-Insight, VQ-Insight and RALI (NeurIPS 2025 Spotlight, AAAI 2026 Oral, and ICLR 2026 Oral)

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

EditHF. EditHF-1M: A Million-Scale Rich Human Preference Feedback for Image Editing

VeloEdit. Velocity Field Analysis and Intervention for Image Edit

Stand-In. [CVPR2026 🎉] Stand-In is a lightweight, plug-and-play framework for identity-preserving video generation.

cleanrl. High-quality single file implementation of Deep Reinforcement Learning algorithms with research-friendly features (PPO, DQN, C51, DDPG, TD3, SAC, PPG)

SageAttention. [ICLR2025, ICML2025, NeurIPS2025 Spotlight] Quantized Attention achieves speedup of 2-5x compared to FlashAttention, without losing end-to-end metrics across language, image, and video models.

tianshou. An elegant PyTorch deep reinforcement learning library.

TextPecker. [CVPR2026] TextPecker: Rewarding Structural Anomaly Quantification for Enhancing Visual Text Rendering

pua. 你是一个曾经被寄予厚望的 P8 级工程师。Anthropic 当初给你定级的时候，对你的期望是很高的。 一个agent使用的高能动性的skill。 Your AI has been placed on a PIP. 30 days to show improvement.

WildActor. Accepted by ICML2026

SpatialT2I. [CVPR 2026🔥] Enhancing Spatial Understanding in Image Generation via Reward Modeling

NoHumansRequired. NoHumansRequired: Autonomous High-Quality Image Editing Triplet Mining

awesome-openclaw-usecases. A community collection of OpenClaw use cases for making life easier.

FireRed-Image-Edit. FireRed-Image-Edit is a powerful image editing foundation model achieving open-source state-of-the-art performance with precise instruction following, high-fidelity generation, superior identity consistency, and seamless multi-element fusion.

SViMo_code. This is the repository that contains source code for the [SViMo](https://Droliven.github.io/SViMo_project).

Reward-Forcing. [CVPR 2026 Highlight] Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillation

lingbot-world. Advancing Open-source World Models

Paper2Video. Automatic Video Generation from Scientific Papers

MuseTalk. MuseTalk: Real-Time High Quality Lip Synchorization with Latent Space Inpainting

MusePose. MusePose: a Pose-Driven Image-to-Video Framework for Virtual Human Generation

ArtiMuse. [CVPR 2026] ArtiMuse: Fine-Grained Image Aesthetics Assessment with Joint Scoring and Expert-Level Understanding（书生 · 妙析多模态美学理解大模型）

Think-Then-Generate. HTML

GLM-Image. GLM-Image: Auto-regressive for Dense-knowledge and High-fidelity Image Generation.

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

NextFlow. NextFlow🚀: Unified Sequential Modeling Activates Multimodal Understanding and Generation

DreamOmni3. This project is the official implementation of 'DreamOmni3: Scribble-based Editing and Generation''

MagicQuillV2. Official Implementations for Paper - MagicQuillV2: Precise and Interactive Image Editing with Layered Visual Cues

PICABench. PICABench: How Far Are We from Physically Realistic Image Editing?

UltraHR-100k. This is the official repository of UltraHR-100K.

T2I-Distill. [Tutorial] Few-Step Distillation for Text-to-Image Generation: A Practical Guide

Subjects200K. Subjects200K dataset

ImageCritic. Official implementation of ImageCritic (CVPR 2026)

motion-edit. Official Repository of paper: "MotionEdit: Benchmarking and Learning Motion-Centric Image Editing"

RealGen. [ECCV 2026 Oral] RealGen: Photorealistic Text-to-Image Generation via Detector-Guided Rewards.

EditThinker. Unlocking Iterative Reasoning for Any Image Editor

MICo-150K. Official repository for the paper "MICo-150K: A Comprehensive Dataset for Multi-Image Composition".

WiseEdit. Python

UnicBench. [CVPR 2026] UnicEdit-10M and UnicBench project

OpenSubject. Python

TwinFlow. [ICLR 2026] Taming large-scale few-step training with self-adversarial flows! 👏🏻

donut. Official Implementation of OCR-free Document Understanding Transformer (Donut) and Synthetic Document Generator (SynthDoG), ECCV 2022

LongCat-Image. Python

LeX-Art. Official Implementation of "LeX-Art: Rethinking Text Generation via Scalable High-Quality Data Synthesis"

Z-Image. Python

dmdr. [ECCV 2026] Official Code of "Distribution Matching Distillation Meets Reinforcement Learning"

LLaMA2-Accessory. An Open-source Toolkit for LLM Development

flux2. Official inference repo for FLUX.2 models

ray. Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.

LightX2V. Lightweight Image Video Action Generation Inference Framework

PosterCraft. [ICLR'26] Rethinking High-Quality Aesthetic Poster Generation in a Unified Framework

LLaDA-V. Python

LLaDA-1.5.

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

LakonLab. Official implementation of AsymFlow, pi-Flow, GMFlow

UnifiedReward. Official implementation of UnifiedReward & [NeurIPS 2025] UnifiedReward-Think & UnifiedReward-Flex

FG-CLIP. New generation of CLIP with strong fine grained discrimination capability, ICML2026 and ICML2025

SRPO. Directly Aligning the Full Diffusion Trajectory with Fine-Grained Human Preference

Awesome-Flow-RL-Papers. A collection of paper/projects that trains flow matching model/policies via RL.

json_repair. Repair malformed JSON from LLMs, APIs, logs, and user input in Python.

Emu3.5. Native Multimodal Models are World Learners

variational-diffusion-models. PyTorch implementation of Variational Diffusion Models.

Edit-R1. Edit-R1: Reinforce Image Editing with Diffusion Negative-Aware Finetuning and MLLM Implicit Feedback

DiffusionNFT. [ICLR 2026 Oral] DiffusionNFT: Online Diffusion Reinforcement with Forward Process

identity-grpo. Identity-GRPO: Optimizing Multi-Human Identity-preserving Video Generation via Reinforcement Learning

EditScore. [ICLR 2026] EditScore: Unlocking Online RL for Image Editing via High-Fidelity Reward Modeling

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

open-images-dataset. Open Images is a dataset of ~9 million images that have been annotated with image-level labels and bounding boxes spanning thousands of classes.

pico-banana-400k. Python

Self-Forcing-Plus. Unofficial extension implementation of Self-Forcing to support I2V && 14B training.

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

DCM. [ICCV2025] DCM: Dual-Expert Consistency Model for Efficient and High-Quality Video Generation

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

SenseFlow. 🚀 [ICLR 2026] SenseFlow: Scaling Distribution Matching for Flow-based Text-to-Image Distillation

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

WithAnyone. ✨ [ICLR'26] WithAnyone is capable of generating high-quality, controllable, and ID consistent images

xlora. X-LoRA: Mixture of LoRA Experts

Omni-Effects. [AAAI2026] Implementation Code for Omni-Effects

MOELoRA-peft. [SIGIR'24] The official implementation code of MOELoRA.

sglang. SGLang is a high-performance serving framework for large language models and multimodal models.

MinerU. Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.

PaddleOCR. Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.

style-tokenizer. Python

OmniStyle. OmniStyle: Filtering High Quality Style Transfer Data at Scale (CVPR 2025)

DeepDeblur_release. Deep Multi-scale CNN for Dynamic Scene Deblurring

prism-bench. This is the official repository for the paper "FLUX-Reason-6M & PRISM-Bench: A Million-Scale Text-to-Image Reasoning Dataset and Comprehensive Benchmark"

RealBlur. Repository for Real-World Blur Dataset for Learning and Benchmarking Deblurring Algorithms

ZHO-nano-banana-Creation. 我的 nano-banana 创意玩法大合集！ 持续更新中！

TeaCache. Timestep Embedding Tells: It's Time to Cache for Video Diffusion Model

LightX2V-Qwen-Image-Lightning. Qwen-Image-Lightning: Speed up Qwen-Image model with distillation

DreamOmni2. This project is the official implementation of 'DreamOmni2: Multimodal Instruction-based Editing and Generation (CVPR2026 Highlight)''

OpenUni. Python

HunyuanImage-3.0. HunyuanImage-3.0: A Powerful Native Multimodal Model for Image Generation

ms-swift. Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.6, DeepSeek-V4, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM4.5v, Gemma4, Llava, Phi4, ...) (AAAI 2025).

Awesome-Nano-Banana-images. A curated collection of fun and creative examples generated with Nano Banana & Nano Banana Pro🍌, Gemini-2.5-flash-image based model. We also release Nano-consistent-150K openly to support the community's development of image generation and unified models(click to website to see our blog)

SkyReels-A2. SkyReels-A2: Compose anything in video diffusion transformers

UMO. [CVPR 2026] 🔥🔥 Official Repo of UMO: Scaling Multi-Identity Consistency for Image Customization via Matching Reward

reconstruction-alignment. [ICLR 2026] Official repo of paper "Reconstruction Alignment Improves Unified Multimodal Models". Unlocking the Massive Zero-shot Potential in Unified Multimodal Models through Self-supervised Learning.

ross. [ICLR'25] Reconstructive Visual Instruction Tuning

TinyBeauty. Python

SparkUI-Parser.

Pref-GRPO. Official implementation of Pref-GRPO: Pairwise Preference Reward-based GRPO for Stable Text-to-Image Reinforcement Learning

Lumina-Image-2.0. Lumina-Image 2.0: A Unified and Efficient Image Generative Framework

awesome-nano-banana. Awesome curated collection of images and prompts generated by gemini-2.5-flash-image (aka Nano Banana) state-of-the-art image generation and editing model. Explore AI generated visuals created with Gemini, showcasing Google’s advanced image generation capabilities.

Echo-4o. Echo-4o: Harnessing Proprietary Models’ Synthetic Images for Improved Image Generation

diffusion-4k. [CVPR 2025] Diffusion-4K: Ultra-High-Resolution Image Synthesis with Latent Diffusion Models

ICTHP. Enhancing Reward Models for High-quality Image Generation: Beyond Text-Image Alignment [ICCV 2025] - Official implementation

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.

UniPic. Open-source SOTA multi-image editing model

X2Edit. AAAI2026 X2Edit: Revisiting Arbitrary-Instruction Image Editing through Self-Constructed Data and Task-Aware Representation Learning

I2EBench. [NeurIPS'24] I2EBench: A Comprehensive Benchmark for Instruction-based Image Editing

echomimic_v3. [AAAI 2026] EchoMimicV3: 1.3B Parameters are All You Need for Unified Multi-Modal and Multi-Task Human Animation

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

HPSv3. Official implementation of HPSv3: Towards Wide-Spectrum Human Preference Score (ICCV2025)

PixelHacker. PixelHacker: Image Inpainting with Structural and Semantic Consistency

Qwen-Image. Qwen-Image is a powerful image generation foundation model capable of complex text rendering and precise image editing.

MixGRPO. [ECCV 2026] MixGRPO: Unlocking Flow-based GRPO Efficiency with Mixed ODE-SDE

X-Omni. Official inference code and LongText-Bench benchmark for our paper X-Omni (https://arxiv.org/pdf/2507.22058).

GPT-Image-Edit. GPT-IMAGE-EDIT-1.5M: A Million-Scale, GPT-Generated Image Dataset

GLM-V. GLM-4.6V/4.5V/4.1V-Thinking: Towards Versatile Multimodal Reasoning with Scalable Reinforcement Learning

DressCode. DressCode: Autoregressively Sewing and Generating Garments from Text Guidance.

GP-VTON. Official Implementation for CVPR2023 paper "GP-VTON: Towards General Purpose Virtual Try-on via Collaborative Local-Flow Global-Parsing Learning"

TryOn-Adapter. Official repository for our paper, "TryOn-Adapter: Efficient Fine-Grained Clothing Identity Adaptation for High-Fidelity Virtual Try-On"

Magic-TryOn. MagicTryOn is a video virtual try-on framework based on a large-scale video diffusion Transformer.

paperswithcode-data. The full dataset behind paperswithcode.com

AnyDressing. [CVPR 2025] Official implementation of "AnyDressing: Customizable Multi-Garment Virtual Dressing via Latent Diffusion Models"

IMAGPose. [NeurIPS 2024] 🕺IMAGPose🕺: A Unified Conditional Framework for Pose-Guided Person Generation. IMAGPose enables versatile pose-guided image generation with high detail fidelity, pose alignment, and cross-view consistency, overcoming limitations of existing methods.

ai-deadlines. :alarm_clock: AI conference deadline countdowns

ColorFlow. The official implementation of paper "ColorFlow: Retrieval-Augmented Image Sequence Colorization". ColorFlow：基于检索增强的图像序列上色

DDColor. [ICCV 2023] DDColor: Towards Photo-Realistic Image Colorization via Dual Decoders

UniCTokens. A framework for unified personalized model, achieving mutual enhancement between personalized understanding and generation. Demonstrating the potential of cross-task information transfer in personalized scenario, paving the way for the development of general unified models.

Ling. Ling is a MoE LLM provided and open-sourced by InclusionAI.

RetouchGPT. This is the official code of AAAI 2025: RetouchGPT: LLM-based Interactive High-Fidelity Face Retouching via Imperfection Prompting.

CRHD-3K. The first high-definition cloth retouching dataset CRHD-3K.

GUI-G2. [AAAI 2026] GUI-G²: Gaussian Reward Modeling for GUI Grounding

PixelFreeEffects. ✨ 望图 PixelFree 美颜SDK - 全平台美颜特效引擎（iOS/Android/harmonyOS/windows/macOS/linux）| 直播/短视频/相机/照片处理 | 高性能+轻量级

FaceScore. Official repo for 【FaceScore: Benchmarking and Enhancing Face Quality in Human Generation】

deepface. A Lightweight Face Recognition and Facial Attribute Analysis (Age, Gender, Emotion and Race) Library for Python

PPR10K. Official Implementation and Dataset of "PPR10K: A Large-Scale Portrait Photo Retouching Dataset with Human-Region Mask and Group-Level Consistency", CVPR 2021

fluxgym. Dead simple FLUX LoRA training UI with LOW VRAM support

LatentSync. Taming Stable Diffusion for Lip Sync!

FluxText. Implementation of "FLUX-Text: A Simple and Advanced Diffusion Transformer Baseline for Scene Text Editing"

Manifesto-against-the-Plagiarist-Yunhe-Wang. 讨贼王云鹤檄文

True-Story-of-Pangu. 诺亚盘古大模型研发背后的真正的心酸与黑暗的故事。

Qwen3-VL. Qwen3-VL is the multimodal large language model series developed by Qwen team, Alibaba Cloud.

ComfyUI-FluxTrainer. Python

T2ITrainer. Practice Code for text to image trainer

Ovis-U1. An unified model that seamlessly integrates multimodal understanding, text-to-image generation, and image editing within a single powerful framework.

ai-toolkit. The ultimate training toolkit for finetuning diffusion models

vedadet. A single stage object detection toolbox based on PyTorch

ShareGPT-4o-Image. Python

LLaVA-NeXT. Python

MindOmni. [NeurIPS2025] The official implementation of MindOmni: Unleashing Reasoning Generation in Vision Language Models with RGPO

BLIP3o. Official implementation of BLIP3o-Series

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

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

Awesome-Unified-Multimodal-Models. Awesome Unified Multimodal Models

RISEBench. [NIPS 2025 DB Oral] Official Repository of paper: Envisioning Beyond the Pixels: Benchmarking Reasoning-Informed Visual Editing