Chanhyuk Lee

self-conditioned-fmlm. Python

posterior-refinement. Official Code Repo for Paper: Posterior Refinement

s-flm. 𝕊-FLM: Language Modeling with Hyperspherical Flows

W-Flow. Official code release for the paper "One-Step Generative Modeling via Wasserstein Gradient Flows"

TinyRecursiveModels. Python

IMDM. Official implementation of "Infinite Mask Diffusion for Few-Step Distillation" (ICML 2026)

JacobiForcing. [ICML 2026] Jacobi Forcing: Fast and Accurate Diffusion-style Decoding

mtp-lm. Source code to accompany research paper on training multi token prediction language models using self-distillation.

Discrete-Diffusion-Forcing. Discrete Diffusion Forcing (D2F): dLLMs Can Do Faster-Than-AR Inference

LRAgent. Official implementation of LRAgent: Efficient KV Cache Sharing for Multi-LoRA LLM Agents

CCDD. Official Repository for ICML 2026 paper Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasoner

MARS. Official Implementation of MARS

ELF-pytorch. Python

DISCO. Code for the DISCO model: General Multimodal Protein Design Enables DNA-Encoding of Chemistry

obsidian-biblib. Obsidian-based reference and literature note manager.

WeDLM. WeDLM: The fastest diffusion language model with standard causal attention and native KV cache compatibility, delivering real speedups over vLLM-optimized baselines.

LangFlow. The first continuous diffusion language model that rivals discrete counterparts on standard language modeling benchmarks like LM1B and OpenWebText.

Claude-Code-Usage-Monitor. Real-time Claude Code usage monitor with predictions and warnings

k-skill. 한국인을 위한 스킬 모음집 - SRT, KTX, 카카오톡, 한글과컴퓨터, 날씨, 미세먼지, 법령, 주식정보, 조선왕조실록, KBO, K-리그, LCK, 특허 검색, 토스 증권, 맞춤법 검사, 중고차 가격, 쿠팡, 네이버 블로그, 다이소, 올리브영, 택배 송장 조회 등등...

flow-divergence-sampler. Jupyter Notebook

marimo. A reactive notebook for Python — run reproducible experiments, query with SQL, execute as a script, deploy as an app, and version with git. Stored as pure Python. All in a modern, AI-native editor.

cmux. Open source Ghostty-based macOS terminal with vertical tabs and notifications for AI coding agents. Built for multitasking, organization, and programmability.

drifting. Python

research-prompts. AI prompts for accelerating the research workflow.

PixelREPA. [ECCV 2026] Official code of "Representation Alignment for Just Image Transformers is not Easier than You Think"

autoresearch. AI agents running research on single-GPU nanochat training automatically

ssd. A lightweight inference engine supporting speculative speculative decoding (SSD).

NAP. Why Diffusion Language Models Struggle with Truly Parallel (Non-Autoregressive) Decoding?

zotflow. Keep your research in flow

distillate. AI research agent that reads papers, runs experiments, and tells you what worked!

obsidian-skills. Agent skills for Obsidian. Teach your agent to use Obsidian CLI and open formats including Markdown, Bases, JSON Canvas.

mcpvault. A lightweight Model Context Protocol (MCP) server for safe Obsidian vault access

mfm. Official Implementation of "Meta Flow Maps enable scalable reward alignment"

ilya-sutskever-recommended-reading. It is said that, Ilya Sutskever gave John Carmack this reading list of ~ 30 research papers on deep learning.

flowbind. We propose an efficient flow-based multimodal generation model with bidirectional flows.

flow-maps. Official codebase for the paper "How to build a consistency model: Learning flow maps via self-distillation" (NeurIPS 2025).

Energy-Diffusion-LLM. Reproduce ICLR2025 Energy-Based Diffusion Language Models for Text Generation

candi-diffusion. CANDI: Continuous and Discrete Diffusion

modified-shortcut-models-pytorch. PyTorch implementation of Shortcut Models [Frans, 2025] with little modification

shortcut-models. Python

jvp_flash_attention. Flash Attention Triton kernel with support for second-order derivatives

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

tvm. Terminal Velocity Matching

flash-attention. Fast and memory-efficient exact attention

MeanFlow. PyTorch implementation of MeanFlow & iMF (one-step generative modeling).

duo. [ICML 2025] The Diffusion Duality

nanoDD. Simple Scalable Discrete Diffusion for text in PyTorch

Fk-Diffusion-Steering. A general framework for inference-time scaling and steering of diffusion models with arbitrary rewards.

research-template. An ML research template with good documentation by Boyuan Chen, an MIT PhD student

mujoco-py. MuJoCo is a physics engine for detailed, efficient rigid body simulations with contacts. mujoco-py allows using MuJoCo from Python 3.

DOGE. Official implementation of "Modeling Multi-Task Model Merging as Adaptive Projective Gradient Descent".

hindsight-experience-replay. This is the pytorch implementation of Hindsight Experience Replay (HER) - Experiment on all fetch robotic environments.

UFC. [NeurIPS 2025] Universal Few-Shot Spatial Control for Diffusion Models

SynthER. Synthetic Experience Replay

pgr. Prioritized Generative Replay (ICLR 2025 Oral)

Awesome-Inference-Time-Scaling. Paper List of Inference/Test Time Scaling/Computing

sioyek. Sioyek is a PDF viewer with a focus on textbooks and research papers

Awesome-Physics-Cognition-based-Video-Generation. A comprehensive list of papers investigating physical cognition in video generation, including papers, codes, and related websites.

llamppl. Probabilistic programming with large language models

diffusion-forcing-transformer. [ICML 2025] Official PyTorch Implementation of "History-Guided Video Diffusion"

Awesome-World-Models. A comprehensive list of papers for the definition of World Models and using World Models for General Video Generation, Embodied AI, and Autonomous Driving, including papers, codes, and related websites.

Awesome-Controllable-Video-Generation. [ArXiv 2025] A survey about controllable video generation: This repo is the official awesome of "Controllable video generation: A survey"

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

diffuser. Code for the paper "Planning with Diffusion for Flexible Behavior Synthesis"

Muon. Muon is an optimizer for hidden layers in neural networks

ctrl-x. Official implementation of "Ctrl-X: Controlling Structure and Appearance for Text-To-Image Generation Without Guidance" (NeurIPS 2024)

ReDi. Official implementation of ReDi: Rectified Discrete Flow (NeurIPS 2025)

perceiver-pytorch. Implementation of Perceiver, General Perception with Iterative Attention, in Pytorch

mup. maximal update parametrization (µP)

cosmos-predict1. Cosmos-Predict1 is a collection of general-purpose world foundation models for Physical AI that can be fine-tuned into customized world models for downstream applications.

uv. An extremely fast Python package and project manager, written in Rust.

git-re-basin. Code release for "Git Re-Basin: Merging Models modulo Permutation Symmetries"

powerlevel10k. A Zsh theme

CART. Python

monaspace. An innovative superfamily of fonts for code

zotero-tab-limiter. Simple plugin for limiting maximum number of open tabs in Zotero.

zotero-plugin-template. A plugin template for Zotero.

Prompt-Diffusion. Official PyTorch implementation of the paper "In-Context Learning Unlocked for Diffusion Models"

evolutionary-model-merge. Official repository of Evolutionary Optimization of Model Merging Recipes

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

awesome-weight-space-learning. Awesome papers on weight-space learning

AdaRank. Official codebase for AdaRank: Adaptive Rank Pruning for Enhanced Model Merging (ICLR 2026)

ImageBind. ImageBind One Embedding Space to Bind Them All

task_singular_vectors. Task Singular Vectors: Reducing Task Interference in Model Merging. Merge models avoiding task interference through separable models.

training_free_model_merging. This repository is the implementation of the paper Training Free Pretrained Model Merging (CVPR2024).

tall_masks. Official repository of "Localizing Task Information for Improved Model Merging and Compression" [ICML 2024]

VLM_survey. Collection of AWESOME vision-language models for vision tasks

Awesome-Model-Merging-Methods-Theories-Applications. Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities. ACM Computing Surveys, 2026.

Awesome-Image-Composition. A curated list of papers, code and resources pertaining to image composition/compositing or object/subject insertion/addition/compositing, which aims to generate realistic composite image.

Awesome-Image-Harmonization. A curated list of papers, code and resources pertaining to image harmonization.

riemannian-fm. code for "Riemannian Flow Matching on General Geometries".

awesome-llm-interpretability. A curated list of Large Language Model (LLM) Interpretability resources.

riemannian-score-sde. Score-based generative models for compact manifolds

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

xatlas. Mesh parameterization / UV unwrapping library

loss-landscape. Code for visualizing the loss landscape of neural nets

supsup. Code for "Supermasks in Superposition"

Text2Tex. [ICCV 2023] Text2Tex: Text-driven Texture Synthesis via Diffusion Models

Awesome-Equivariant-Frame.

riemannian-diffusion-mixture-torch. PyTorch implementation for "Generative Modeling on Manifolds Through Mixture of Riemannian Diffusion Processes" (ICML 2024).

simulation-free-node. Official implementation of the paper: Simulation-Free Training of Neural ODEs on Paired Data (NeurIPS 2024)

ViT-Prisma. ViT Prisma is a mechanistic interpretability library for Vision and Video Transformers (ViTs).

AdaMerging. AdaMerging: Adaptive Model Merging for Multi-Task Learning. ICLR, 2024.

weight-ensembling_MoE. Code for paper "Merging Multi-Task Models via Weight-Ensembling Mixture of Experts"

Twin-Merging. [NeurIPS2024] Twin-Merging: Dynamic Integration of Modular Expertise in Model Merging

Diffusion-Assignment4-Distillation. Python

notero. A Zotero plugin for syncing items and notes into Notion

peta. Code for paper "Parameter Efficient Multi-task Model Fusion with Partial Linearization"

task_vectors. Editing Models with Task Arithmetic

tangent_task_arithmetic. Source code of "Task arithmetic in the tangent space: Improved editing of pre-trained models".

Awesome-Incremental-Learning. Awesome Incremental Learning

Diffusion-Assignment1-DDPM. Jupyter Notebook

CS322_fall2024. KAIST. Formal Languages and Automata. Fall semester.

awesome. 😎 Awesome lists about all kinds of interesting topics

awesome-deep-phenomena. A curated list of papers of interesting empirical study and insight on deep learning. Continually updating...

cli. Command-line tool to customize Spotify client. Supports Windows, macOS, and Linux.

vllm. A high-throughput and memory-efficient inference and serving engine for LLMs

UPop. [ICML 2023] UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers

QDrop. The official PyTorch implementation of the ICLR2022 paper, QDrop: Randomly Dropping Quantization for Extremely Low-bit Post-Training Quantization

apex. A PyTorch Extension: Tools for easy mixed precision and distributed training in Pytorch

CLIP. CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image

Awesome-Quantization-Papers. List of papers related to neural network quantization in recent AI conferences and journals.

llm_slerp_generation. Repo hosting codes and materials related to speeding LLMs' inference using token merging.

Awesome-Knowledge-Distillation. Awesome Knowledge-Distillation. 分类整理的知识蒸馏paper(2014-2021)。

Awesome-Dataset-Distillation. A curated list of awesome papers on dataset distillation and related applications.

MinimaxDiffusion. [CVPR2024] Efficient Dataset Distillation via Minimax Diffusion

mtt-distillation. Official code for our CVPR '22 paper "Dataset Distillation by Matching Training Trajectories"