This is your work, valued
The only thing we have to fear is fear itself - nameless, unreasoning, unjustified terror which paralyzes needed efforts to convert retreat into advance.
stgcn. The PyTorch implementation of STGCN.
★ 694awesome-gnn. A list for GNNs and related works.
★ 110chebnet. The PyTorch version of ChebNet.
★ 39cautious_adam. The PyTorch implementation of Cautious-Adam.
★ 18tiger. A Tight-fisted Optimizer (Tiger), implemented in PyTorch.
★ 12converter. The official PyTorch implementation of Converter.
★ 7dcrnn_data. The traffic data files for Los Angeles (METR-LA) and the Bay Area (PEMS-BAY).
★ 7adanorm. The PyTorch implementation of AdaNorm.
★ 3GCN. The PyTorch version of GCN implemented by the paper.
★ 3mgc_v1. Magnetic Graph Convolutional Networks v1
★ 3chebgibbsnet. The official PyG implementation of ChebGibbsNet.
★ 3mgc_v2. Magnetic Graph Convolutional Network v2
★ 3rmsnorm. The PyTorch implementation of RMSNorm.
★ 2GDC. The PyTorch version of GDC implemented by the paper.
★ 2awesome-ssm. A list for SSMs and related works.
★ 2SGC. The PyTorch version of SGC implemented by the paper.
★ 1SSGC. The PyTorch version of SSGC implemented by the paper.
★ 1bgu. The PyTorch implementation of Binary Gating Unit.
★ 1early_stopping. The PyTorch implementation of early stopping.
★ 1DiGCN. The PyTorch version of DiGCN implemented by the paper.
★ 1HKGCN. The PyTorch version of HKGCN implemented by the paper.
★ 1uninstally. A clean, native macOS uninstaller. Completely removes apps and every file they leave behind using smart bundle-identifier detection. SwiftUI + Finder extension.
★ 300optimizerDuck. Free, open-source Windows optimization tool for performance, privacy, and simplicity.
★ 6.1kopen-science. Open Science Desktop — local-first, model-agnostic AI research workbench for macOS, Windows & Linux. Open-source Claude Science desktop alternative built on Tauri + MCP + agent skills.
★ 626lobehub. 🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team.
★ 80ksparse-delta-memory. This repositories contains the reference implementation for the Sparse Delta Memory paper.More precisely, it contains the model definition as well as triton and cuda kernels for the Sparse Delta Memory layer.
★ 20wordfreq. Access a database of word frequencies, in various natural languages.
★ 1.7knvidia-kaggle. NVIDIA Kaggle Plugin gives agents end-to-end Kaggle competition workflows through a single skill, nvidia-kaggle-skill. It can gather competition context, study public writeups and notebooks, reproduce kernels locally, submit to competitions, and manage Ka
★ 241docker_images_sync. docker镜像一键拉取,无需梯子/服务器,无数量限制docker images sync workflow and scripts
★ 317Pip-Helper. Pip-Helper(A simple and easy-to-use browser extension that helps you watch videos in picture-in-picture mode on any webpage.)
★ 84sumatrapdf-plus. C
★ 371all-routes-lead-to-collapse. A cross-architecture study of routing concentration in transformers, state-space models, recurrent networks, and graphs.
★ 2Free-NTFS-for-Mac. Nigate: An open-source NTFS utility for Mac. It supports all Mac models (Intel and Apple Silicon), providing full read-write access, mounting, and management for NTFS drives.
★ 2.2kjlens-qwen36. J-space / Jacobian-lens visualizer for Qwen3.6-27B (4-bit) on Apple Silicon, ported to Apple MLX
★ 302jacobian-lens. Companion code for the global workspace interpretability paper
★ 1.1kSpotifly. Native Spotify client
★ 126posterly. Build academic conference posters as a single HTML/CSS file, rendered to print-ready PDF via headless Chromium. A coding-agent skill.
★ 274delta-attention-residuals-code. Delta Attention Residuals - supplementary code and pretrained models
★ 39FluentCleaner. A real WinUI desktop cleaner with terminal control and multi-database support
★ 3kOnlyofficePersonal. ONLYOFFICE Personal 是一款基于wasm运行的Office办公套件,无需服务端直接运行。
★ 273residual-stream-duality. Official Project Page for Residual Stream Duality in Modern Transformer Architectures (https://arxiv.org/abs/2603.16039)
★ 7OpenLiveWriter. An open source fork of Windows Live Writer
★ 2.5kopen-r1. Fully open reproduction of DeepSeek-R1
★ 26klm-evaluation-harness. A framework for few-shot evaluation of language models.
★ 13kai-model-fingerprint. AI 模型随机数分布指纹数据集 — 17 模型 / OpenRouter / 含原始响应与相似度分析
★ 64WeChatTweak. A command-line tool for tweaking WeChat - 首款微信 macOS 客户端撤回拦截与多开 🔨
★ 14korca. Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop and mobile.
★ 16kazurelinux. General purpose Linux OS for Azure
★ 5.3kwloc. 修改 Apple 网络定位(gs-loc)返回坐标 · 支持 Surge / Quantumult X / Loon / Stash · 快捷指令一键设置/恢复定位
★ 3.9kVeraCrypt. Disk encryption with strong security based on TrueCrypt
★ 11kUn-0. Un-0: an image generator powered by a simulated system of coupled oscillators, an example of an emerging physical computing substrate.
★ 301DeepSpeed. DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.
★ 43kAirBattery. Get the battery usage of all your devices on your Mac and put them on the Dock / Menu Bar / Widget! && 在Mac上获取你所有设备的电量信息并显示在Dock / 状态栏 / 小组件上!
★ 2.9krascal. Swift
★ 120Asspp. The App Store for your multi-account eco system.
★ 5.7kMegatron-LM. Ongoing research training transformer models at scale
★ 17komni-macos. Native macOS semantic search over your local files - text, images, audio, video in one vector space, on-device on Apple silicon.
★ 188kidox.app. A Modern Launchpad Replacement for Mac
★ 30flash-muon. Flash-Muon: An Efficient Implementation of Muon Optimizer
★ 257SiameseNorm. Code of the research paper "SiameseNorm: Breaking the Barrier to Reconciling Pre/Post-Norm"
★ 11torchtitan. A PyTorch native platform for training generative AI models
★ 5.5kHDLA-Impl. New version of HDLA's Triton Implementations.
★ 2hc-stream-collapse. Analyzing Stream Collapse in Hyper-Connections (ICML 2026 WSS Workshop)
★ 2LockIME. A native macOS menu-bar app that keeps your keyboard input source locked — global, per-app, and per-URL.
★ 447headroom. Compress tool outputs, logs, files, and RAG chunks before they reach the LLM. 60-95% fewer tokens, same answers. Library, proxy, MCP server.
★ 58kProbMoE. Python
★ 4nanoLlama. A simple implementation of Llama.
★ 10nanollama. Train Llama 3 models from scratch. Any scale, any personality. By Arianna Method.
★ 48llama-moe. ⛷️ LLaMA-MoE: Building Mixture-of-Experts from LLaMA with Continual Pre-training (EMNLP 2024)
★ 1kRMoE. Official implementation of RMoE (Layerwise Recurrent Router for Mixture-of-Experts)
★ 33megablocks. Python
★ 1.6kReMoE. [ICLR2025] Codebase for "ReMoE: Fully Differentiable Mixture-of-Experts with ReLU Routing", built on Megatron-LM.
★ 118Autonomy-of-Experts. Python
★ 10glm-coding. Python
★ 661tex2word. LaTeX 2 Word XML
★ 45ANE. Training neural networks on Apple Neural Engine via reverse-engineered private APIs
★ 6.9kforel. File automation for macOS. The open source Hazel alternative for macOS.
★ 248IPA-Download. 一款用于安装 IPA 历史版本的工具,适用于获取旧版应用并自动捕获数据包。下载后,可直接通过 AirDrop 传输至 iPhone、iPad 上并安装并使用。
★ 1.5kflash-linear-attention-feedstock. A conda-smithy repository for flash-linear-attention.
★ 1grit. Reimplementing Git in Rust. Like a boss.
★ 484deepseek-pp. DeepSeek Web browser extension: AI agent workspace with MCP tools, memory, Skills, automation, web search, and conversation export.
★ 1.2kBark. Bark is an iOS App which allows you to push custom notifications to your iPhone
★ 8.6kenableMacosAI. 国行 Mac 一键开启完整 Apple 智能(端侧 + Private Cloud Compute 云端)· macOS 27 / Apple Silicon
★ 1.7klm-engine. LM engine is a library for pretraining/finetuning LLMs
★ 182MiMo-Code. MiMo Code: Where Models and Agents Co-Evolve
★ 12kMSA. Python
★ 373modded-nanogpt. NanoGPT (124M) in 90 seconds
★ 5.5kno-mistakes. git push no-mistakes
★ 5.9kthoughts-memo-translation. Thoughts Memo 汉化组翻译合集
★ 166smt. Pretraining Recurrent Networks without Recurrence
★ 59api-relay-audit. Local security audit for AI API relays and LLM proxies: detects prompt injection, model substitution, tool-call rewriting, SSE anomalies, error leakage, and Web3 wallet risks.
★ 752metacubexd. Mihomo Dashboard, The Official One, XD
★ 4kRoutingFreeMoE. Official repository for paper Routing-Free Mixture-of-Experts.
★ 12nanoLlama. nanoGPT reproduction but for llama 2
★ 10SlimBrave-Neo. SlimBrave Neo — Debloat and harden Brave Browser on Linux and Windows. Python curses TUI + CLI. Zero dependencies.
★ 493litgpt. 20+ high-performance LLMs with recipes to pretrain, finetune and deploy at scale.
★ 13kSamba. [ICLR 2025] Samba: Simple Hybrid State Space Models for Efficient Unlimited Context Language Modeling
★ 964supermium. Chromium fork for Windows XP/2003 and up
★ 4.6kSimpleGPT. [ICML 2026] Improving GPT via a simple normalization strategy
★ 15opendataloader-pdf. PDF Parser for AI-ready data. Automate PDF accessibility. Open-source.
★ 27kwall-attention-release. Attention variant with per-channel multiplicative decay
★ 47OpenLogi. ⚡️A native, local-first alternative to Logitech Options+, written in Rust 🦀 — remap buttons, DPI, and SmartShift over HID++. No account, no telemetry.
★ 6.1kMouser. A lightweight, open-source, fully local alternative to Logitech Options+ for remapping Logitech HID++ mice.
★ 4.8kThaw. Menu bar manager for macOS 26
★ 8.5kkimi-agent-rs. Kimi Agent (Rust), the Kimi Code CLI Wire mode-compatible agent server.
★ 3Parallax. Official repository for Parallax (Parameterized Local Linear Attention)
★ 63nanochat. The best ChatGPT that $100 can buy.
★ 54wasn-style. A WASN-style LaTeX template for paper preprints.
★ 1Selene-TV. 以 MoonTV v100 版本 / Helios 为后端的 Android TV 客户端,针对遥控器设备优化
★ 532geist-font. HTML
★ 3.5kinter. The Inter font family
★ 20kfree-font. 大概是2020年最全的免费可商用字体,这里收录的商免字体都能找到明确的授权出处,可以放心使用,持续更新中...
★ 6.4kTikTokSans. A free and open-source font by TikTok
★ 206fonts. Font files available from ByteDance Fonts
★ 321muonNC-llama135m.
★ 1mozilla-type-family. A collection of custom fonts for Mozilla
★ 177plex. The package of IBM’s typeface, IBM Plex.
★ 11kncu-style. An NCU-style LaTeX template for paper preprints.
★ 1ml-intern. 🤗 ml-intern: an open-source ML engineer that reads papers, trains models, and ships ML models
★ 11kamuse. AMUSE optimizer implementation
★ 33freellmapi. OpenAI-compatible proxy that stacks the free tiers of 16 LLM providers (~1.7B tokens/month) behind one /v1 endpoint — plus any custom OpenAI-compatible endpoint. Smart routing, automatic failover, encrypted keys. Personal experimentation only.
★ 16kKun. AI agent workspace with Code Write and Design modes built into your application.
★ 5.2kDeepSeek-Reasonix. DeepSeek-native AI coding agent for your terminal. Engineered around prefix-cache stability — leave it running.
★ 27kGatedDeltaNet-2. Official PyTorch Implementation of Gated DeltaNet-2: Decoupling Erase and Write in Linear Attention
★ 237SympFormer. A new transformer architecture based on principles from accelerated gradient flows and their *symp*lectic discretizations
★ 2ngpt. Normalized Transformer (nGPT)
★ 209swift-markdown-engine. A native AppKit Markdown editor for macOS, built on TextKit 2 and bridged to SwiftUI.
★ 727TrustTunnel. Modern, fast and obfuscated VPN protocol
★ 3.4kxsched. A scheduling framework for multitasking over diverse XPUs, including GPUs, NPUs, ASICs, and FPGAs
★ 175motrix-next. A full-featured download manager — rebuilt from the ground up
★ 8.6kSolaar. Linux device manager for Logitech devices
★ 9kcontra-muon. Python
★ 5aurora-release. Aurora optimizer release
★ 150Riemannian-Residual-Neural-Networks. [NeurIPS 2023] Riemannian Residual Neural Networks (https://arxiv.org/abs/2006.10254)
★ 25how-to-train-your-gpt. Build a modern LLM from scratch. Every line commented. Explained like we are five.
★ 2.3kStableSPAM. Python
★ 28minimind. 🧠「大模型」2小时完全从0训练64M的小参数LLM!Train a 64M-parameter LLM from scratch in just 2h!
★ 53kmHC-manifold-constrained-hyper-connections. implementations and experimentation on mHC by deepseek - https://arxiv.org/abs/2512.24880
★ 367hyper-connections. Attempt to make multiple residual streams from Bytedance's Hyper-Connections paper accessible to the public
★ 186flash-sinkhorn. The official repository of FlashSinkhorn [ICML 2026 Oral]
★ 204MasterDnsVPN. Advanced DNS tunneling VPN for censorship bypass, optimized beyond DNSTT and SlipStream with low-overhead ARQ, resolver load balancing, high packet-loss stability and speed.
★ 6.8kMano-Restriking-Manifold-Optimization-for-LLM-Training. The official code of "Mano: Restriking Manifold Optimization for LLM Training".
★ 25FlashQLA. high-performance linear attention kernel library built on TileLang
★ 584FlashMLA. FlashMLA: Efficient Multi-head Latent Attention Kernels
★ 13kPearcleaner. A free, source-available and fair-code licensed mac app cleaner
★ 14kflame. 🔥 A minimal training framework for scaling FLA models
★ 401SwiftVLC. A Swift wrapper around libVLC for iOS, macOS, tvOS, visionOS, and Mac Catalyst
★ 524AirPrint_Bridge. AirPrint Bridge: Enable AirPrint for Unsupported Printers on macOS
★ 595PolarExpress. Python
★ 32uBlock. JavaScript
★ 30preconditioned-deltanet. Shell
★ 4softpick-attention. Landing repository for the paper "Softpick: No Attention Sink, No Massive Activations with Rectified Softmax"
★ 92Notepads. A modern, lightweight text editor with a minimalist design.
★ 10kseqmdl-design. Python
★ 1sniffnet. Comfortably monitor your Internet traffic 🕵️♂️
★ 40kTileKernels. A kernel library written in tilelang
★ 1.6ktriattention. TriAttention — Efficient long reasoning with trigonometric KV cache compression. Enables OpenClaw local deployment on memory-constrained GPUs.
★ 811FlashKDA. FlashKDA: high-performance Kimi Delta Attention kernels
★ 450GatedDeltaNet. [ICLR 2025] Official PyTorch Implementation of Gated Delta Networks: Improving Mamba2 with Delta Rule
★ 619deltanet. An implementation of the paper "Parallelizing Linear Transformers with the Delta Rule over Sequence Length"
★ 7Keychron-Keyboards-Hardware-Design. Industrial design files for Keychron keyboards and mice. 100+ models with CAD assets in STEP, DXF, DWG, and PDF. Source-available, with commercial use allowed for original compatible accessories within the license terms.
★ 3.5khermes-agent. The agent that grows with you
★ 213kHLA. Official Project Page for HLA: Higher-order Linear Attention (https://arxiv.org/abs/2510.27258)
★ 102tuning_playbook. A playbook for systematically maximizing the performance of deep learning models.
★ 30kShichiZip. The 7-Zip derivative intended for macOS
★ 882EUGen. Source code for EUGens: Efficient, Unified and General Dense Layers
★ 1Recordly. Create polished demo videos without editing skills. Mac/Windows/Linux
★ 19kIPQuality. IP质量检测脚本 - IP Quality Check Script
★ 9.5kedgetunnel. edgetunnel2 VLESS/Trojan/SS 多功能面板
★ 40kvox. Voice input reimagined — speak in any language, type in any language.
★ 108legado-for-mac-pub. 这是 Legado(开源阅读)的 macOS 版本,使用 Swift 和 SwiftUI 构建。Legado支持的书源、订阅,都可以平滑地从安卓中迁移过来,在【导入书源】后【解析】,基本实现了【在线找书】-> 【阅读】 -> 【换源】 和 【订阅】 -> 【解析】超方便的一体化工具链。并且在阅读逻辑上,基本保持了人类的阅读习惯。
★ 58free-code. The free build of Claude Code. All telemetry removed, security-prompt guardrails stripped, all experimental features enabled.
★ 8.6kopenscreen. Create stunning demos for free. Open-source, no subscriptions, no watermarks, and free for commercial use. An alternative to Screen Studio.
★ 39kneko-master. A modern and elegant dashboard for network traffic visualization and analysis.
★ 2.2kcuLA. CUDA kernels for linear attention variants, written in CuTe DSL and CUTLASS C++.
★ 528TTC-Net. [ICML'26] Beyond Test-Time Memory: State-Space Optimal Control for LLM Reasoning
★ 15WindInput. 清风输入法 | 轻量、快速、可定制的开源中文输入法(支持Windows/macOS)| 项目正在进行 Rust 的重构,当前版本不再更新。
★ 146claude-code-sourcemap. TypeScript
★ 9.5klearn-coding-agent. Research on Coding Agents
★ 12kclaurst. Agentic Coding for Builders who Ship
★ 10kclaw-code. An agent-managed museum exhibit, built in Rust with Gajae-Code / LazyCodex — developed and maintained with no human intervention.
★ 195kcc-gateway. AI API identity gateway — reverse proxy that normalizes device fingerprints and telemetry for privacy-preserving API proxying
★ 3kgram-newton-schulz. Fast Polar Decomposition for Muon
★ 163PasteMemo. PasteMemo - Smart clipboard manager for macOS
★ 108jmlr-style-file. LaTeX style file for the Journal of Machine Learning Research
★ 174RaBitQ. The repo has been moved to https://github.com/VectorDB-NTU/RaBitQ-Library. [SIGMOD 2024] RaBitQ: Quantizing High-Dimensional Vectors with a Theoretical Error Bound for Approximate Nearest Neighbor Search
★ 252MagicPods-Windows. Add little magic to your Airpods
★ 838type4me. MacOS语音输入法,实时识别、大模型文本优化、全本地存储
★ 1.3kMarkFlowy. AI-powered cross-platform editor for WYSIWYG Markdown and general-purpose text editing.
★ 2.3kSelf-MoE. The official pre-view of CS-MoE.
★ 6awesome-unix. All the UNIX and UNIX-Like: Linux, BSD, macOS, Illumos, 9front, and more.
★ 1.8kreynard-browser. An experimental Gecko-based web browser for iOS 13+.
★ 985koe. A zero-GUI macOS voice input tool. Press a hotkey, speak, and the corrected text is pasted into whatever app you're using.
★ 352app-store-preflight-skills. AI agent skill to scan iOS/macOS projects for App Store rejection patterns before submission
★ 1.3kMSA. Memory Sparse Attention - A scalable, end-to-end trainable latent-memory framework for 100M-token contexts.
★ 3.5kchatwise-releases. The fastest AI Chatbot for any LLM
★ 1.4kExplorerPatcher. This project aims to enhance the working environment on Windows
★ 33kAttention-Residuals.
★ 3.3khelium-windows. Helium Browser for Windows
★ 1.2klogi-options-plus-mini. A minimal Logi Options+ deployment preset for keyboard and mouse only, with configurable disabling of optional features like Logi Voice, Flow, Analytics, and more
★ 1.6kPixiv-SwiftUI. 为现代 Apple 平台设计的第三方 Pixiv 客户端,支持 iOS/iPadOS/macOS
★ 73koharu. ML-powered manga translator, written in Rust.
★ 4.9kgemini-voyager. An all-in-one enhancement suite for Google Gemini & AI Studio - timeline navigation, folder management, prompt library, and chat export in one powerful extension. / Google Gemini & AI Studio 全能增强插件:集成时间轴导航、文件夹管理、提示词库及聊天导出等众多功能。
★ 19khelium-macos. Helium Browser for macOS
★ 1.1kHTmuon. Python
★ 15rat-plus. RAT+: Train Dense, Infer Sparse - Recurrence Augmented Attention for Dilated Inference (ICML2026)
★ 10browser. Lightpanda: the headless browser designed for AI and automation
★ 32kbest-windows-apps. 推荐好用、优秀的 Windows 应用
★ 3.9kLLMmap. Python
★ 371MetaLA. Offical implementation of "MetaLA: Unified Optimal Linear Approximation to Softmax Attention Map" (NeurIPS2024 Oral)
★ 36omlx. LLM inference server with continuous batching & SSD caching for Apple Silicon — managed from the macOS menu bar
★ 18kChromeAppHeroes. 🌈谷粒-Chrome插件英雄榜, 为优秀的Chrome插件写一本中文说明书, 让Chrome插件英雄们造福人类~ ChromePluginHeroes, Write a Chinese manual for the excellent Chrome plugin, let the Chrome plugin heroes benefit the human~ 公众号「0加1」同步更新
★ 26kflashoptim. Python
★ 256nanoGPT. The simplest, fastest repository for training/finetuning medium-sized GPTs.
★ 61kCAN. Python
★ 173twitter-cli. A CLI for Twitter/X — feed, bookmarks, and user timeline in terminal
★ 2.7kDeepEP. DeepEP: an efficient expert-parallel communication library
★ 9.8kscion. Python
★ 68TorchCode. 🔥 LeetCode for PyTorch — practice implementing softmax, attention, GPT-2 and more from scratch with instant auto-grading. Jupyter-based, self-hosted or try online.
★ 4.3kpicc. Rust macOS automation toolkit — screenshot, OCR, accessibility CLI, break reminder, and more, built on Apple native frameworks via objc2
★ 98EchoMusic. 🎉 一个简约的第三方酷狗概念版音乐播放器
★ 2.3k