Starred repositories
进京证 CLI 工具 - 让人类和 AI Agent 都能在终端中办理进京证 | npm install -g cross-bj
Gemma 4 26B-A4B inference in ~2 GB of RAM on any M-series MacBook
《深入理解 AI Infra：量化分析与系统设计》（李博杰 著）开源书稿：从硬件约束和模型架构出发，量化推导 LLM 推理与训练系统设计。含全书正文、PDF、配套计算工具与实验
Run 284B DeepSeek-V4-Flash (<59GiB RAM: 41.1 cold pp, 9.1 N=1 tg, up to 25.3 spec tok/s) & 30B MoEs (22.4 N=1 / 111.6 spec tok/s) on 64GB CPU (0 GPUs).
DeepSeek Harness Agent Workflow
A moddable training framework for pre-training ablations.
HelixHarness 提供完整的 AI Coding 工程化基础能力，还支持不同企业和团队结合自身的业务知识、工程规范和研发流程，持续构建和演进属于自己的专属 Harness
Offline, privacy-first grammar checker. Fast, open-source, Rust-powered
Art and animation, written as code. Illustrations, loops, interactive web art, launch-videos and scored films in dozens of styles, identical on every render.
墨韵 · A sheet of rice paper that breathes: real-time GPU ink fluid, a pentatonic guqin, and an invisible hand that paints a new landscape every time.
Real-time, photoreal shallow water in a single HTML file. WebGL2, no libraries, no build step.
Cross-vendor 3D Gaussian Splatting trainer - video to splat to mesh, Vulkan or CUDA.
AG-UI: the Agent-User Interaction Protocol. Bring Agents into Frontend Applications.
Open-source AI coworkers that each get a computer of their own: a browser, files and tools, with every action decided before it happens and recorded after. Bring any AG-UI agent.
Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per re…
Orca lets you run a fleet of coding agents (e.g. Claude Code, Codex) with your own subscription. Available on desktop, mobile and remote runtime.
A local inference engine for Apple silicon, built around the model.
Let AI agents use your real, logged-in browser without interrupting your work. CLI + extension for browser automation across any shell-capable AI agent.
Make humans and AI agents work as one team — open-source and self-hostable.
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
微信读书（WeRead）增强工具包：集成 Agent、API、SDK、CLI 与插件，内置 40 项操作；支持阅读微信公众号、书架与笔记读写、书籍导入及 AI 权限。
Frontier-class open models on a free Kaggle TPU v5e-8: GLM-5.3-Flash 320B MoE (~165 tok/s, our own JAX engine) and Qwen3.8-27B bf16 (~130 tok/s), 262k context, prefix caching. Works with Claude Cod…
Ongoing research training transformer models at scale
Miles is an enterprise-facing reinforcement learning framework for LLM and VLM post-training, forked from and co-evolving with slime.
Assignments for CS146S: The Modern Software Dev (Stanford University Fall 2026/2025)