Fangyikang Wang

Megatron-LM. Ongoing research training transformer models at scale

Audio-Omni. [SIGGRAPH 2026] Repository of Audio-Omni

alphaflow. Official pytorch implementation of "AlphaFlow: Understanding and Improving MeanFlow Models"

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

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）

Gaussian-Shading. [CVPR 2024] Gaussian Shading: Provable Performance-Lossless Image Watermarking for Diffusion Models

tree-ring-watermark. Python

deepgen. A Lightweight Unified Multimodal Model for Advancing Image Generation and Editing

awesome-ai-research-writing. Elevate your AI research writing, no more tedious polishing ✨

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.

ai-for-mathematics-and-optimization-slides. Python

BitDance. BitDance & UniWeTok: Open-source autoregressive model with binary visual tokens. A research project for building powerful multimodal autoregressive model.

imeanflow. Official Implementation of iMF https://arxiv.org/abs/2512.02012

SEAL. [ICCV 2025] Official implementation of 'SEAL: Semantic Aware Image Watermarking'

Kimi-Linear.

efficient-tune. Python

hugo-theme-saas-landing-page. 🚀 Ship a SaaS landing page this afternoon. Hero, pricing, testimonials — AI generates pages, you own as Markdown. No React to maintain, free hosting forever. 创业落地页，AI 一键生成 👇

WeTok. [ICLR2026] WeTok: Powerful Discrete Tokenization for High-Fidelity Visual Reconstruction

al-folio. A beautiful, simple, clean, and responsive Jekyll theme for academics

transformer. Transformer: PyTorch Implementation of "Attention Is All You Need"

academic-poster-template-pptx. A simple academic poster template using pptx, for NeurIPS/ICML/ICCV.

LML-diffusion-sampler. [ICCV 2025] Official implementation of "Unleashing High-Quality Image Generation in Diffusion Sampling Using Second-Order Levenberg-Marquardt-Langevin".

DiffusionFisher. [ICML 2025] Official implementation of "Efficiently Access Diffusion Fisher: Within the Outer Product Span Space".

Video-GPT. [ICLR2026] Video-GPT via Next Clip Diffusion.

DPMs-with-Adam. This is the code for "Boosting Diffusion Models with an Adaptive Momentum Sampler"

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

FramePack. Lets make video diffusion practical!

Phased-Consistency-Model. [NeurIPS 2024] Boosting the performance of consistency models with PCM!

fkc-diffusion. Jupyter Notebook

one-small-step. 这是一个简单的技术科普教程项目，主要聚焦于解释一些有趣的，前沿的技术概念和原理。每篇文章都力求在 5 分钟内阅读完成。

PickScore. Python

AdjointDEIS. [NeurIPS 2024] Official PyTorch implementation for the paper "AdjointDEIS: Efficient Gradients for Diffusion Models"

TextToucher. Official Pytorch Implementation for "TextToucher: Fine-Grained Text-to-Touch Generation" (AAAI 2025)

BELM. [NeurIPS 2024] Official implementation of "BELM: Bidirectional Explicit Linear Multi-step Sampler for Exact Inversion in Diffusion Models".

diff-sampler. An open-source toolbox for fast sampling of diffusion models. Official implementations of our works published in ICML, NeurIPS, CVPR, J. Stat. Mech.

UniPC. [NeurIPS 2023] UniPC: A Unified Predictor-Corrector Framework for Fast Sampling of Diffusion Models

Discrete-Probability-Flow. [NeurIPS 2023] Formulating Discrete Probability Flow Through Optimal Transport

DriveDiTFit. Official Pytorch Implementation for "DriveDiTFit: Fine-tuning Diffusion Transformers for Autonomous Driving Data Generation" (TOMM)

Reversi-AI-based-on-Monte-Carlo-tree-search. 基于蒙特卡洛树搜索算法编写的黑白棋AI算法

KiNet-lean. A simplified implementation of KiNet with support for only linear Fokker-Planck

mle_score_ode. Official code for "Maximum Likelihood Training for Score-Based Diffusion ODEs by High-Order Denoising Score Matching" (ICML 2022)

DiffAN. Diffusion Models for Causal Discovery

i-DODE. Official code for "Improved Techniques for Maximum Likelihood Estimation for Diffusion ODEs" (ICML 2023)

HPSv2. Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences of Text-to-Image Synthesis

zjuthesis. Zhejiang University Graduation Thesis LaTeX Template

PnPInversion. [ICLR2024] Official repo for paper "PnP Inversion: Boosting Diffusion-based Editing with 3 Lines of Code"

improved-aesthetic-predictor. CLIP+MLP Aesthetic Score Predictor

openai-cookbook. Examples and guides for using the OpenAI API

AdjointDPM. Python

diffusers. 🤗 Diffusers: State-of-the-art diffusion models for image, video, and audio generation in PyTorch.

PNDM. The official implementation for Pseudo Numerical Methods for Diffusion Models on Manifolds (PNDM, PLMS | ICLR2022)

EDICT. Jupyter Notebook

BDIA. Python

CGProject. C++

SKM. Covariate Shift in Hilbert Space: A Solution via Surrogate Kernels

variational_wgf. [ICML2022] Variational Wasserstein gradient flow

JDOT. Joint distribution optimal transportation for domain adaptation

swf. TeX

guided-diffusion. Python

EPT. The official implementation of Deep Generative Learning via Euler Particle Transport.

academicpages.github.io. Github Pages template based upon HTML and Markdown for personal, portfolio-based websites.

Deep-Reinforcement-Learning-Hands-On-Second-Edition. Deep-Reinforcement-Learning-Hands-On-Second-Edition, published by Packt

DQN-robot-arm. Deep Q learning algorithm written on PyTorch for solving 2D robot arm reacher

sinkhorn-imitation. Code for reproducing the experiment results of the paper Imitation Learning with Sinkhorn Distances.

mirror-stein-samplers. Sampling with Mirrored Stein Operators

MIED. Source code for "Sampling with Mollified Interaction Energy Descent" [ICLR 2023].

cvxbook_additional_exercises. Additional exercises and data for EE364a. No solutions; for public consumption.

Robot-Maze-Solving. 机器人走迷宫，Pytorch，强化学习，DQN。

zju-se-project. 浙江大学 2022 春夏《软件工程》课程期末作业。要求 5 人为一小组负责一个模块，5 个模块组成一个完整的医疗管理系统。

zju-icicles. 浙江大学课程攻略共享计划

os-lab. 基于RISC-V的仿Linux操作系统内核