仓库/Skill 库

从零开始一步步用 PyTorch 实现一个推理 LLM。Implement a reasoning LLM in PyTorch from scratch, step by step

Transformer 模型的高速推理引擎。Fast inference engine for Transformer models

RLHF 精选资源列表（持续更新）。A curated list of reinforcement learning with human feedback resources (continually updated)

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)

TypeDB：为系统而生，而非为记录而生。TypeDB: Built for systems, not records

终端中的 Agent，配备本地工具：编写代码、使用终端、浏览网页。在其之上构建你自己的持久化自主 Agent。Your agent in your terminal, equipped with local tools: writes code, uses the terminal, browses the web. Make your own persistent autonomous agent on top!

免费学习如何使用 LLMOps 最佳实践构建端到端生产级 LLM & RAG 系统：源码 + 12 个实操课程。🤖 𝗟𝗲𝗮𝗿𝗻 for 𝗳𝗿𝗲𝗲 how to 𝗯𝘂𝗶𝗹𝗱 an end-to-end 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝗶𝗼𝗻-𝗿𝗲𝗮𝗱𝘆 𝗟𝗟𝗠 & 𝗥𝗔𝗚 𝘀𝘆𝘀𝘁𝗲𝗺 using 𝗟𝗟𝗠𝗢𝗽𝘀 best practices: ~ 𝘴𝘰𝘶𝘳𝘤𝘦 𝘤𝘰𝘥𝘦 + 12 𝘩𝘢𝘯𝘥𝘴-𝘰𝘯 𝘭𝘦𝘴𝘴𝘰𝘯𝘴

面向 AI Agent 的开发者控制平面Developer Control Plane for your AI Agents

一个覆盖文本、图像、视频与音频任务的统一多模态评估工具包。One-for-All Multimodal Evaluation Toolkit Across Text, Image, Video, and Audio Tasks

大型多模态模型 LMM 的开源评估工具包，支持 220+ LMM 与 80+ 基准测试。Open-source evaluation toolkit of large multi-modality models (LMMs), support 220+ LMMs, 80+ benchmarks

【三年面试五年模拟】AIGC/LLM/AI Agent算法工程师面试秘籍。涵盖AIGC、LLM大模型、AI Agent、具身智能、传统深度学习、自动驾驶、机器学习、计算机视觉、自然语言处理、强化学习、大数据挖掘、世界模型、元宇宙、AGI等AI行业面试笔试干货经验与核心知识。

开源 LLMOps 平台：集成 prompt playground、prompt 管理、LLM 评估和 LLM 可观测性。The open-source LLMOps platform: prompt playground, prompt management, LLM evaluation, and LLM observability all in one place.

为 LLM 优化推理代理Optimizing inference proxy for LLMs

CSGHub 是 OpenCSG 团队开发的全新开源 LLM 管理平台，提供开源与本地化/SaaS 解决方案，功能对标 Hugging Face。兼容 Hugging Face Python SDK，全面掌控 LLM、数据集和 agent 的全生命周期。CSGHub is a brand-new open-source platform for managing LLMs, developed by the OpenCSG team. It offers both open-source and on-premise/SaaS solutions, with features comparable to Hugging Face. Gain full control over the lifecycle of LLMs, datasets, and agents, with Python SDK compatibility with Hugging Face. Join us! ⭐️

针对加速基础设施和微服务架构优化的生成式 AI 参考工作流。Generative AI reference workflows optimized for accelerated infrastructure and microservice architecture.

在 Kubernetes 上借助现代加速器实现 SOTA 推理性能。Achieve state of the art inference performance with modern accelerators on Kubernetes

ChatGPT、Claude 等 LLM 的所有前端 GUI 客户端汇总。Every front-end GUI client for ChatGPT, Claude, and other LLMs

任意模型，任意硬件，零妥协。基于 @ziglang / @openxla / MLIR / @bazelbuild 构建。Any model. Any hardware. Zero compromise. Built with @ziglang / @openxla / MLIR / @bazelbuild

用于加速视频生成的统一推理与后训练框架。A unified inference and post-training framework for accelerated video generation.

可扩展至上千个微调 LLM 的 Multi-LoRA 推理服务器Multi-LoRA inference server that scales to 1000s of fine-tuned LLMs

基于 PaddlePaddle 的高性能 LLM 与 VLM 推理部署工具包。High-performance Inference and Deployment Toolkit for LLMs and VLMs based on PaddlePaddle

作为记忆层的 Agent SkillsAgent Skills as a Memory Layer

⚡FlashRAG：面向高效 RAG 研究的 Python 工具包（WWW2025 Resource）⚡FlashRAG: A Python Toolkit for Efficient RAG Research (WWW2025 Resource)

面向 LLM 实验和 AI Agent 的评估与追踪Evaluation and Tracking for LLM Experiments and AI Agents

面向 LLM 评估和 AI agent 测试的平台The platform for LLM evaluations and AI agent testing

🚀 通过易用的硬件优化工具，加速 🤗 Transformers、Diffusers、TIMM 和 Sentence Transformers 的推理与训练。🚀 Accelerate inference and training of 🤗 Transformers, Diffusers, TIMM and Sentence Transformers with easy to use hardware optimization tools

Apple Silicon 上最快的本地 AI 引擎。比 Ollama 快 4.2 倍，缓存 TTFT 仅 0.08s，工具调用支持率 100%。内置 17 种工具解析器、prompt cache、推理分离、云端路由。可作为 OpenAI 的即插即用替代，兼容 Claude Code、Cursor、Aider。The fastest local AI engine for Apple Silicon. 4.2x faster than Ollama, 0.08s cached TTFT, 100% tool calling. 17 tool parsers, prompt cache, reasoning separation, cloud routing. Drop-in OpenAI replacement. Works with Claude Code, Cursor, Aider.

ReMe：面向 Agent 的记忆管理工具包——Remember Me, Refine Me.ReMe: Memory Management Kit for Agents - Remember Me, Refine Me.

🕹️ 一个面向开发者的开源 LLMOps 平台，旨在简化 prompt 设计、版本管理、即时交付、协作、排障与可观测性等全流程。🕹️ Open-source, developer-first LLMOps platform designed to streamline prompt design, version management, instant delivery, collaboration, troubleshooting, observability and more.

AGiXT 是一个动态 AI Agent 自动化平台，无缝编排跨多个 AI provider 的指令管理和复杂任务执行，结合自适应记忆、智能特性与可扩展插件系统，提供高效全面的 AI 解决方案。AGiXT is a dynamic AI Agent Automation Platform that seamlessly orchestrates instruction management and complex task execution across diverse AI providers. Combining adaptive memory, smart features, and a versatile plugin system, AGiXT delivers efficient and comprehensive AI solutions.

LLM 交互的安全工具包The Security Toolkit for LLM Interactions

📚 OpenVINO™ 的 Jupyter notebook 教程。📚 Jupyter notebook tutorials for OpenVINO™

Laminar —— 专为 AI agent 打造的开源可观测性平台，YC S24。Laminar - open-source observability platform purpose-built for AI agents. YC S24.

用于对 LLM prompt 进行对抗测试的开源可视化编程环境An open-source visual programming environment for battle-testing prompts to LLMs.

面向编码 Agent 的通用 Playwright 自动化。General-purpose Playwright automation for coding agents

PyTorch 原生的量化和稀疏化方案，支持训练与推理PyTorch native quantization and sparsity for training and inference