wnma

deepcode-cli. Deep Code 是专为 deepseek-v4 模型优化的终端 AI 编码助手，支持深度思考、推理强度控制以及 Agent Skills。

multi-agent-mlx. Python

DFT. [ICLR 2026] On the Generalization of SFT: A Reinforcement Learning Perspective with Reward Rectification.

WeightWatcher. The WeightWatcher tool for predicting the accuracy of Deep Neural Networks

multica. The open-source managed agents platform. Turn coding agents into real teammates — assign tasks, track progress, compound skills.

Liger-Kernel. Efficient Triton Kernels for LLM Training

hyperframes. Write HTML. Render video. Built for agents.

agent-yes. Run AI coding agents (Claude, Codex, Gemini …) unattended — auto-answer prompts, auto-retry on rate limits, and list/tail/steer every agent locally or from agent-yes.com

ao. PyTorch native quantization and sparsity for training and inference

torchtitan. A PyTorch native platform for training generative AI models

auto-arch-tournament. VHDL

superpowers. An agentic skills framework & software development methodology that works.

reader. Convert any URL to an LLM-friendly input with a simple prefix https://r.jina.ai/

sub2api. Sub2API 一站式开源中转服务，让 Claude、Openai 、Gemini、Grok订阅统一接入，支持拼车共享，更高效分摊成本，原生工具无缝使用。

dflash. DFlash: Block Diffusion for Flash Speculative Decoding

BlackHole. BlackHole is a modern macOS audio loopback driver that allows applications to pass audio to other applications with zero additional latency.

trl. Train transformer language models with reinforcement learning.

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

qiushi-skill. 求是Skill——从经典唯物辩证法与实践哲学中提炼出一条总原则和九大方法论工具武装AI大脑。Qiushi-Skill: Build agents that investigate first, focus on the main contradiction, validate in practice, and keep pushing until the work is actually done.

frontend-slides. Create beautiful slides on the web using a coding agent's frontend skills

parameter-golf. Train the smallest LM you can that fits in 16MB. Best model wins!

RandOpt. Official Codebase for "Neural Thickets: Diverse Task Experts Are Dense Around Pretrained Weights" (ICML 2026 Spotlight)

agency-agents. A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.

StockApiPython. 歪枣网提供股票数据接口，股票数据采集，基金数据接口，股票level2行情接口，Lv2数据，实时行情接口，股票行情数据，股票量化数据，股票交易数据，k线行情数据，股票概念数据，K线、TICK、L2逐笔成交港股美股CTP金融高频行情历史数据等数据服务。

happy. Mobile and Web client for Codex and Claude Code, with realtime voice, encryption and fully featured

monty. A minimal, secure Python interpreter written in Rust for use by AI

llm_distillation_playbook. Best practices for distilling large language models.

gemini-cli. An open-source AI agent that brings the power of Gemini directly into your terminal.

biliup. 自动直播录制、投稿、twitch、ytb频道搬运工具。命令行投稿(B站)和视频下载工具，提供多种登录方式，支持多p。

biliup-rs. 命令行投稿(B站)和视频下载工具，提供多种登录方式，支持多p。

siyuan. A privacy-first, self-hosted, fully open source personal knowledge management software, written in typescript and golang.

Soft-Thinking. Official implementation of the NeurIPS 2025 paper "Soft Thinking: Unlocking the Reasoning Potential of LLMs in Continuous Concept Space"

RankSurprisalRatio. [ACL 2026 Main] Official Repo for Paper "Which Reasoning Trajectories Teach Students to Reason Better? A Simple Metric of Informative Alignment“

openclaw. Your own personal AI assistant. Any OS. Any Platform. The lobster way. 🦞

Multiplex-Thinking. Multiplex Thinking: Reasoning via Token-wise Branch-and-Merge

notebooklm-py. Unofficial Python API and agentic skill for Google NotebookLM. Full programmatic access to NotebookLM's features—including capabilities the web UI doesn't expose—via Python, CLI, and AI agents like Claude Code, Codex, and OpenClaw.

DroPE. Extending the Context of Pretrained LLMs by Dropping Their Positional Embedding

shuowen. 說文解字的檢索數據

witr. Why is this running?

NEFTune. Official repository of NEFTune: Noisy Embeddings Improves Instruction Finetuning

neko. A self hosted virtual browser that runs in docker and uses WebRTC.

Profiles4limbo. 毒奶自用，懒人配置文件（Quantumult X）：去广告分流规则、Tiktok解锁重写、VSCO解锁、神机分流、blackmatrix7分流规则。

BLEUnlock. Lock/unlock your Mac with your iPhone, Apple Watch, or any other Bluetooth LE devices

KittenTTS. State-of-the-art TTS model under 25MB 😻

macOCR. Get any text on your screen into your clipboard.

vision-parse. Parse PDFs into markdown using Vision LLMs

index-tts. An Industrial-Level Controllable and Efficient Zero-Shot Text-To-Speech System

Dayflow. The automatic work journal/time tracker. Privately turns your screen into a timeline of what you actually accomplished. Open-source and local-first.

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)

NVSHMEM-Tutorial. NVSHMEM‑Tutorial: Build a DeepEP‑like GPU Buffer

MoneyPrinterTurbo. 利用AI大模型，一键生成高清短视频 Generate short videos with one click using AI LLM.

prompt-lookup-decoding. Simple speculative decoding technique, integrated in vLLM and transformers

RLHF_in_notebooks. RLHF (Supervised fine-tuning, reward model, and PPO) step-by-step in 3 Jupyter notebooks

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

mirage. Mirage Persistent Kernel: Compiling LLMs into a MegaKernel

tiny-diffusion. A minimal PyTorch implementation of probabilistic diffusion models for 2D datasets.

nano-vllm. Nano vLLM

drawnix. 开源白板工具（SaaS），一体化白板，包含思维导图、流程图、自由画等。All in one open-source whiteboard tool with mind, flowchart, freehand and etc.

tensor.h. creating a tiny tensor library in raw C

podcastfy. An Open Source Python alternative to NotebookLM's podcast feature: Transforming Multimodal Content into Captivating Multilingual Audio Conversations with GenAI

Seed1.5-VL. Seed1.5-VL, a vision-language foundation model designed to advance general-purpose multimodal understanding and reasoning, achieving state-of-the-art performance on 38 out of 60 public benchmarks.

OpenInTerminal. ✨ Finder Toolbar app for macOS to open the current directory in Terminal, iTerm, Hyper or Alacritty.

EAGLE. Official Implementation of EAGLE-1 (ICML'24), EAGLE-2 (EMNLP'24), and EAGLE-3 (NeurIPS'25).

LookaheadDecoding. [ICML 2024] Break the Sequential Dependency of LLM Inference Using Lookahead Decoding

Paddle. PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice （『飞桨』核心框架，深度学习&机器学习高性能单机、分布式训练和跨平台部署）

zotero-pdf2zh. PDF2zh for Zotero | Zotero PDF中文翻译插件

llama3.np. llama3.np is a pure NumPy implementation for Llama 3 model.

deepwiki-open. Open Source DeepWiki: AI-Powered Wiki Generator for GitHub/Gitlab/Bitbucket Repositories. Join the discord: https://discord.gg/gMwThUMeme

DeepEP. DeepEP: an efficient expert-parallel communication library

excalidraw. Virtual whiteboard for sketching hand-drawn like diagrams

gitignore. A collection of useful .gitignore templates

mlx-vlm. MLX-VLM is a package for inference and fine-tuning of Vision Language Models (VLMs) on your Mac using MLX.

nsz. NSZ - Homebrew compatible NSP/XCI compressor/decompressor

motioneye. A web frontend for the motion daemon.

zotero-gpt. GPT Meet Zotero.

OpenDiloco. OpenDiLoCo: An Open-Source Framework for Globally Distributed Low-Communication Training

zerox. OCR & Document Extraction using vision models

Kokoro-FastAPI. Dockerized FastAPI wrapper for Kokoro-82M text-to-speech model w/multiplatform CPU, AMD, NVIDIA GPU PyTorch support, handling, and auto-stitching

LayerSkip. Code for "LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding", ACL 2024

pytorch-to-safetensor-converter. A simple converter which converts pytorch bin files to safetensor, intended to be used for LLM conversion.

lupine. LUPINE is a GPU over IP bridge allowing GPUs on remote machines to be attached to CPU-only machines.

RingLink. implemention of RingLink network core function

moonlight-qt. GameStream client for PCs (Windows, Mac, Linux, and Steam Link)

PDFMathTranslate. [EMNLP 2025 Demo] PDF scientific paper translation with preserved formats - 基于 AI 完整保留排版的 PDF 文档全文双语翻译，支持 Google/DeepL/Ollama/OpenAI 等服务，提供 CLI/GUI/MCP/Docker/Zotero

Sunshine. Self-hosted game stream host for Moonlight.

moonlight-android. GameStream client for Android

mflux. MLX native implementations of state-of-the-art generative image models

Mooncake. Mooncake is the serving platform for Kimi, a leading LLM service provided by Moonshot AI.

genesis. Gensis is a lightweight deep learning framework written from scratch in Python, with Triton as its backend for high-performance computing.

pyreft. Stanford NLP Python library for Representation Finetuning (ReFT)

LLMFarm. llama and other large language models on iOS and MacOS offline using GGML library.

Clean-Architecture-zh. 《架构整洁之道》中文翻译

VideoLingo. Netflix-level subtitle cutting, translation, alignment, and even dubbing - one-click fully automated AI video subtitle team | Netflix级字幕切割、翻译、对齐、甚至加上配音，一键全自动视频搬运AI字幕组

SponsorBlock. Skip YouTube video sponsors (browser extension)

LaTeX-OCR. pix2tex: Using a ViT to convert images of equations into LaTeX code.

hivemind. Decentralized deep learning in PyTorch. Built to train models on thousands of volunteers across the world.

Kolors. Kolors Team

dclm. DataComp for Language Models

devzat. The devs are over here at devzat, chat over SSH!

ScreenBroadcaster. Screen broadcast and remote control tool based on FFmpeg. 基于 FFmpeg 的屏幕广播与远程操控工具.

mlx-swift. Swift API for MLX

distributed-llama. Distributed LLM inference. Connect home devices into a powerful cluster to accelerate LLM inference. More devices means faster inference.

anything-llm. Stop renting your intelligence. Own it with AnythingLLM. Everything you need for a powerful local-first agent experience

unsloth. Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

mlc. MiniCPM on Android platform.

the-super-tiny-compiler. :snowman: Possibly the smallest compiler ever

AFFiNE. There can be more than Notion and Miro. AFFiNE(pronounced [ə‘fain]) is a next-gen knowledge base that brings planning, sorting and creating all together. Privacy first, open-source, customizable and ready to use.

AppFlowy. Bring projects, wikis, and teams together with AI. AppFlowy is the AI collaborative workspace where you achieve more without losing control of your data. The leading open source Notion alternative.

docker-webdav. Docker image for running an Apache WebDAV server

vaultwarden. Unofficial Bitwarden compatible server written in Rust, formerly known as bitwarden_rs

mergekit. Tools for merging pretrained large language models.

hyperdiv. Build reactive web UIs in Python

Umi-OCR. OCR software, free and offline. 开源、免费的离线OCR软件。支持截屏/批量导入图片，PDF文档识别，排除水印/页眉页脚，扫描/生成二维码。内置多国语言库。

jepa. PyTorch code and models for V-JEPA self-supervised learning from video.

reor. Private & local AI personal knowledge management app for high entropy people.

gistpad. 📘 VS Code extension for managing and sharing code snippets, notes and interactive samples using GitHub Gists

ml-aim. This repository provides the code and model checkpoints for AIMv1 and AIMv2 research projects.

LLaVA. [NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.

UEVR. Universal Unreal Engine VR Mod (4.8 - 5.4)

UForm. Pocket-Sized Multimodal AI for content understanding and generation across multilingual texts, images, and 🔜 video, up to 5x faster than OpenAI CLIP and LLaVA 🖼️ & 🖋️

ml-ferret. Python

LLMLingua. [EMNLP'23, ACL'24] To speed up LLMs' inference and enhance LLM's perceive of key information, compress the prompt and KV-Cache, which achieves up to 20x compression with minimal performance loss.

JARVIS. Your own personal voice assistant: Voice to Text to LLM to Speech, displayed in a web interface

pgweb. Cross-platform client for PostgreSQL databases

wandb. The AI developer platform. Use Weights & Biases to train and fine-tune models, and manage models from experimentation to production.

mlx. MLX: An array framework for Apple silicon

gpt-fast. Simple and efficient pytorch-native transformer text generation in <1000 LOC of python.

kernl. Kernl lets you run PyTorch transformer models several times faster on GPU with a single line of code, and is designed to be easily hackable.

MiTemperature2. Read the values of the Xiaomi Mi Bluetooth Temperature sensor 2 including custom encrypted format.

1L-Sparse-Autoencoder. Python

influence-release. Jupyter Notebook

TTS. 🐸💬 - a deep learning toolkit for Text-to-Speech, battle-tested in research and production

EMO. [ICLR 2024]EMO: Earth Mover Distance Optimization for Auto-Regressive Language Modeling(https://arxiv.org/abs/2310.04691)

outlines. Structured Outputs

llama2.c. Llama 2 Everywhere (L2E)

droidcam-linux-client. GNU/Linux/nix client for DroidCam

petals. 🌸 Run LLMs at home, BitTorrent-style. Fine-tuning and inference up to 10x faster than offloading

random_quantize. a novel data augmentation method across data modalities

tokenmonster. Ungreedy subword tokenizer and vocabulary trainer for Python, Go, C++ & Javascript

BitFit. Simple Parameter-efficient Fine-tuning for Transformer-based Masked Language-models

selfcheckgpt. SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models

few-shot-meta-baseline. Meta-Baseline: Exploring Simple Meta-Learning for Few-Shot Learning, in ICCV 2021

memit. Mass-editing thousands of facts into a transformer memory (ICLR 2023)

Synergy-Binaries. Download the latest stable Synergy binaries.

rfs. Python

AWGIM. Code for paper "Attentive Weights Generation for Few Shot Learning via Information Maximization"

BF3S. Boosting Few-Shot Visual Learning with Self-Supervision

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

bfs. A breadth-first version of the UNIX find command

cuda_practice. My own repository containing the codes I wrote to practice CUDA programming.

attention_with_linear_biases. Code for the ALiBi method for transformer language models (ICLR 2022)

Queryable. Run OpenAI's CLIP and Apple's MobileCLIP model on iOS to search photos.

NeurIPS2022_UOT_fine_tuning. Python

stats. macOS system monitor in your menu bar

honest_llama. Inference-Time Intervention: Eliciting Truthful Answers from a Language Model

FSKD. Python

audiocraft. Audiocraft is a library for audio processing and generation with deep learning. It features the state-of-the-art EnCodec audio compressor / tokenizer, along with MusicGen, a simple and controllable music generation LM with textual and melodic conditioning.

MAZE. Implementation of the paper "MAZE: Data-Free Model Stealing Attack Using Zeroth-Order Gradient Estimation".

KegNet. Knowledge Extraction with No Observable Data (NeurIPS 2019)

datafree-model-extraction. CVPR 2021 Official repository for the Data-Free Model Extraction paper. https://arxiv.org/abs/2011.14779

CompRess. Compressing Representations for Self-Supervised Learning

MeZO. [NeurIPS 2023] MeZO: Fine-Tuning Language Models with Just Forward Passes. https://arxiv.org/abs/2305.17333

retraining-free-pruning. [NeurIPS 2022] A Fast Post-Training Pruning Framework for Transformers

SWD. Selective Weight Decay to prune continuously neural networks.

neural-nets-are-reversible. Implementation of paper "Deep Neural Networks are Surprisingly Reversible"

mediamtx. Ready-to-use Media-over-QUIC / SRT / WebRTC / RTSP / RTMP / LL-HLS / MPEG-TS / RTP live media server and media proxy that allows to read, publish, proxy, record and playback real-time video and audio streams.

ebook-GPT-translator. Enjoy reading with your favorite style.

egos-2000. Envision a future where everyone can read all the code of an educational operating system.

CTKD. [AAAI 2023] Official PyTorch Code for "Curriculum Temperature for Knowledge Distillation"

PaLM. An open-source implementation of Google's PaLM models

ImageBind. ImageBind One Embedding Space to Bind Them All

study-music. An "awesome music theory" kinda wiki with books, resources and courses for studying everything about music and sound

FanControl.Releases. This is the release repository for Fan Control, a highly customizable fan controlling software for Windows.

slackdump. Save or export your private and public Slack messages, threads, files, and users locally without admin privileges.

sish. HTTP(S)/WS(S)/TCP Tunnels to localhost using only SSH.

Kavita. Kavita is a fast, feature rich, cross platform reading server. Built with the goal of being a full solution for all your reading needs. Setup your own server and share your reading collection with your friends and family.

bark. 🔊 Text-Prompted Generative Audio Model

wasm-gpt. Tensor library for machine learning

awesome-selfhosted. A list of Free Software network services and web applications which can be hosted on your own servers

linuxwave. Generate music from the entropy of Linux 🐧🎵

RedPajama-Data. The RedPajama-Data repository contains code for preparing large datasets for training large language models.

koodo-reader. A modern ebook manager and reader with sync and backup capacities for Windows, macOS, Linux, Android, iOS and Web

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

gitea. Git with a cup of tea! Painless self-hosted all-in-one software development service, including Git hosting, code review, team collaboration, package registry and CI/CD

segment-anything. The repository provides code for running inference with the SegmentAnything Model (SAM), links for downloading the trained model checkpoints, and example notebooks that show how to use the model.

the-algorithm. Source code for the X Recommendation Algorithm

bitsandbytes. Accessible large language models via k-bit quantization for PyTorch.

fairscale. PyTorch extensions for high performance and large scale training.

audioFlux. A library for audio and music analysis, feature extraction.

peft. 🤗 PEFT: State-of-the-art Parameter-Efficient Fine-Tuning.

mkcert. A simple zero-config tool to make locally trusted development certificates with any names you'd like.

rpi-eth-display. Turn a RPi into an additionnal video output over Ethernet

self-instruct. Aligning pretrained language models with instruction data generated by themselves.

Forward-Forward. Reimplementation of Geoffrey Hinton's Forward-Forward Algorithm

optimate. A collection of libraries to optimise AI model performances

ggwave. Tiny data-over-sound library

wave-share. Serverless, peer-to-peer, local file sharing through sound

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

whisper.cpp. Port of OpenAI's Whisper model in C/C++

xla. A machine learning compiler for GPUs, CPUs, and ML accelerators