仓库/Skill 库

Transformers：面向文本、视觉、音频及多模态 SOTA 机器学习模型的模型定义框架，同时支持推理与训练。🤗 Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

潜空间文本到图像扩散模型。A latent text-to-image diffusion model

开源超级 AI 助手与 Agent 编排框架。任务规划、工具与 Skill 调用、基于记忆和知识的自演化。多模型、多通道、轻量、可扩展，一行安装。（原名 chatgpt-on-wechat）Open-source super AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-model, multi-channel. Lightweight, extensible, one-line install. (formerly chatgpt-on-wechat)

🔊 文本提示生成式音频模型🔊 Text-Prompted Generative Audio Model

FinGPT:开源金融大语言模型!革命性 🔥 已在 HuggingFace 发布训练好的模型。FinGPT: Open-Source Financial Large Language Models! Revolutionize 🔥 We release the trained model on HuggingFace.

中文LLaMA&Alpaca大语言模型+本地CPU/GPU训练部署 (Chinese LLaMA & Alpaca LLMs)

面向 agents、助手和企业搜索的私有 AI 平台，内置 Agent Builder、深度研究、文档分析、多模型支持，以及为 agents 提供 API 连接能力。Private AI platform for agents, assistants and enterprise search. Built-in Agent Builder, Deep research, Document analysis, Multi-model support, and API connectivity for agents.

💡 面向语义搜索、LLM 编排与语言模型工作流的一体化 AI 框架💡 All-in-one AI framework for semantic search, LLM orchestration and language model workflows

复旦大学开源的工具增强对话语言模型An open-source tool-augmented conversational language model from Fudan University

面向本地部署的高速 LLM 服务High-speed Large Language Model Serving for Local Deployment

提供 AI 应用和模型服务的最简方式 —— 构建模型推理 API、任务队列、LLM 应用、多模型 pipeline 等。The easiest way to serve AI apps and models - Build Model Inference APIs, Job queues, LLM apps, Multi-model pipelines, and more!

LTX-2 音视频生成模型的官方 Python 推理与 LoRA 训练包。Official Python inference and LoRA trainer package for the LTX-2 audio–video generative model.

中文LLaMA-2 & Alpaca-2大模型二期项目 + 64K超长上下文模型 (Chinese LLaMA-2 & Alpaca-2 LLMs with 64K long context models)

面向全模态模型的高效推理框架。A framework for efficient model inference with omni-modality models

用于高性能 AI 模型服务（vLLM、SGLang）和按需 SSH 访问 GPU 实例的 GPU 集群管理器。A GPU cluster manager for high-performance AI model serving (vLLM, SGLang) and on-demand SSH-accessible GPU instances.

高质量且极速的预训练深度学习模型与 demo。Pre-trained Deep Learning models and demos (high quality and extremely fast)

SOTA 低比特 LLM 量化（INT8/FP8/MXFP8/INT4/MXFP4/NVFP4）与稀疏化方案；面向 PyTorch、TensorFlow 与 ONNX Runtime 的领先模型压缩技术SOTA low-bit LLM quantization (INT8/FP8/MXFP8/INT4/MXFP4/NVFP4) & sparsity; leading model compression techniques on PyTorch, TensorFlow, and ONNX Runtime

面向终端的开源编码 Agent，由社区集体而非公司构建。自带模型，代码保留在本地，不欠任何人。An open coding agent for your terminal, built by a community collective rather than a company. Bring your own model, keep your code on your machine, and owe nothing to anyone.

潜入 Shell 的幽灵。Ante 是一个自包含的 Agent 框架，核心高度优化。体验类似 Claude Code 或 Codex，但无其依赖与模型限制。Ghost in your shell. Ante is a self-contained agent harness with a highly optimized core. It works like Claude Code or Codex, with none of their dependencies or model constraints.

开源 LLM/VLM 负载均衡器与服务平台，用于规模化自托管 LLM（和 VLM）🏓🦙 作为 llm-d、Docker Model Runner 等项目的替代方案，组件更少、部署更简单，基于 ggml 生态构建。支持 CPU 和 GPU。Open-source LLM/VLM load balancer and serving platform for self-hosting LLMs (and VLMs) at scale 🏓🦙 Alternative to projects like llm-d, Docker Model Runner, etc but with less moving parts and simple deployments built around ggml ecosystem. Runs on CPU and GPU.

[GenAI 应用开发框架] 🚀 快速轻松构建 GenAI 应用 💬 在代码中使用结构化数据和链式调用语法与 GenAI Agent 交互 🧩 使用事件驱动的 *TriggerFlow* 管理复杂的 GenAI 业务逻辑 🔀 无需重写代码即可切换任意模型[GenAI Application Development Framework] 🚀 Build GenAI application quick and easy 💬 Easy to interact with GenAI agent in code using structure data and chained-calls syntax 🧩 Use Event-Driven Flow *TriggerFlow* to manage complex GenAI working logic 🔀 Switch to any model without rewrite application code

通过统一的基于 SQL 的框架，为人类和 AI Agent 提供 Cloud、SaaS、API 和 Model Context Protocol (MCP) 资源的查询、配给与运维能力。Query, provision and operate Cloud, SaaS, API and Model Context Protocol (MCP) resources through a unified SQL-based framework for humans and AI agents.

论文"LAMBDA: A large Model Based Data Agent"的官方仓库。https://www.polyu.edu.hk/ama/cmfai/lambda.htmlThis is the offical repository of paper "LAMBDA: A large Model Based Data Agent". https://www.polyu.edu.hk/ama/cmfai/lambda.html

Pax 是一个基于 Jax 的机器学习框架，用于训练大规模模型。Pax 支持先进且完全可配置的实验与并行化，并已展现出业界领先的模型 FLOP 利用率。Pax is a Jax-based machine learning framework for training large scale models. Pax allows for advanced and fully configurable experimentation and parallelization, and has demonstrated industry leading model flop utilization rates.

AI 编码 Agent 的开源控制平面。运行一支 AI 工程团队：并行的 Claude Code、Codex 与 Gemini worker，配备共享看板、原子任务、调度、循环、来源标记消息、模型切换与自愈恢复。一个仪表盘，或一部手机即可掌控。MIT 协议，单个 Rust 二进制。Open-source control plane for AI coding agents. Run an AI engineering team: parallel Claude Code, Codex, and Gemini workers with a shared board, atomic tasks, schedules, loops, origin-stamped messaging, model switching, and self-healing recovery. One dashboard, or your phone. MIT, single Rust binary.

Smithers 是一个 Agent 工作流框架，用于在简单的 TypeScript 配置文件中定义工作流，并快速、持久且可靠地执行它们。Smithers is an agentic workflow framework for defining workflows in simple TypeScript configuration files and executing them quickly, durably, and reliably

DeepSeek Harness 审批请求的 second-model AI 自动复核：只读复核子 Agent 返回结构化 allow/deny 判定及理由，默认 fail-closed，会话日志（approval/asked → autoReview/verdict → approval/decided）全程可审计。Second-model AI auto-review for DeepSeek Harness approval requests: a read-only reviewer subagent returns structured allow/deny verdicts with reasons, fail-closed by default, fully auditable from the session log (approval/asked -> autoReview/verdict -> approval/decided).

开源核心的 AI 工作台——覆盖任何模型的 notebooks、Agent、RAG、语音和图像：OpenAI、Anthropic、Google、xAI，或通过 Ollama/vLLM 的本地模型。BSL 1.1，两年后自动转换为 Apache-2.0。当别人的 AI 停止运行时，你的 AI 仍在运行。The open-core AI workbench — notebooks, agents, RAG, voice, and images across any model: OpenAI, Anthropic, Google, xAI, or local via Ollama/vLLM. BSL 1.1, auto-converting to Apache-2.0 on a two-year clock. Your AI keeps running when theirs doesn't.

🧠💻 将互联网重新构想为自组织思维导图 🤖🔎 STREAM：基于顶部结果提取与答案模型的搜索 📈📝 REASON 文档撰写 Agent 🚜📜 Tractor 文本提取器 🔤📊 SEEKTOPIC🧠💻 Reimagine the Internet as Self-Organizing Mind Map 🤖🔎 STREAM: Search with Top Result Extraction & Answer Model 📈📝 REASON Docs Writing Agent 🚜📜 Tractor the Text Extractor 🔤📊 SEEKTOPIC

发布 AI Agent 制品，而非 demo 表演 — 验证器把关、跨模型族运行，自动学习哪个模型最优。Ship AI-agent artifacts, not demo theater — verifier-gated, cross-family runs that learn which model wins.

[CoLM 2026] MoANT 官方代码：面向多任务大语言模型微调的语义感知秩一专家混合模型[CoLM 2026] Official code for MoANT: Mixture-of-Rank-One-Experts with semantic-aware Intuition for Multi-task Large Language Model Finetuning

面向 AI agent 的基于证据的评估——将每条断言与 agent 真实工具输出进行核对（受约束、基于证据的模型判断，而非整体式 LLM 评判的猜测），并附带置信区间。Evidence-grounded evaluation for AI agents — verifies each claim against the agent's real tool outputs (constrained, evidence-grounded model judgment, not holistic LLM-judge guesswork), with confidence intervals.

邮件的大脑：MailFathom 将 IMAP 邮箱转变为自托管、AI-native 的服务。邮件同步至你自己的 PostgreSQL，建立索引以支持搜索与检索，并通过 Model Context Protocol 提供给 AI agent。具备语义检索、问答与受控的写入工具。基于 .NET 10，AGPL-3.0-only 许可。A brain for your mail: MailFathom turns IMAP mailboxes into a self-hosted, AI-native service. Mail synchronizes into your own PostgreSQL, is indexed for search and retrieval, and is served to AI agents over the Model Context Protocol. Semantic retrieval, answering, and gated write tools. .NET 10, AGPL-3.0-only.

本地 Windows 桌面监视器，用于查看正在运行的 Claude Code agent —— 实时状态、预估成本、模型、主机以及一键聚焦，按项目分组。只读且完全离线。Local Windows desktop monitor for your running Claude Code agents - live status, estimated cost, model, host and one-click focus, grouped by project. Read-only and fully offline.

SharpAI 是基于 llama.cpp（通过 LlamaSharp）构建的可嵌入 Embedding、补全与模型管理平台，内置 Ollama 兼容的 Web 服务。SharpAI is an embeddable embeddings, completions, and model management platform using llama.cpp via LlamaSharp, with a built-in Ollama-compatible webserver.

运行 Ollama 本地 LLM 服务的 Docker 镜像。默认安全，所有 API 请求需 Bearer token（首次启动时自动生成）。OpenAI 兼容 API。支持首次启动模型预拉取、NVIDIA GPU (CUDA) 加速和持久化模型存储。多架构：amd64、arm64。Docker image to run an Ollama local LLM server. Secure by default, all API requests require a Bearer token (auto-generated on first start). OpenAI-compatible API. Supports first-start model pre-pull, NVIDIA GPU (CUDA) acceleration, and persistent model storage. Multi-arch: amd64, arm64.