PhiCookBook — це всеохопний репозиторій кулінарних рецептів, що містить практичні приклади, навчальні посібники та документацію для роботи з сімейством малих мовних моделей (SLMs) Phi від Microsoft. Репозиторій демонструє різноманітні варіанти використання, включно з інференсом, донавчанням, квантизацією, реалізаціями RAG та мультимодальними застосунками на різних платформах і фреймворках. Ключові технології: - Мови: Python, C#/.NET, JavaScript/Node.js - Фреймворки: ONNX Runtime, PyTorch, Transformers, MLX, OpenVINO, Semantic Kernel - Платформи: Microsoft Foundry, GitHub Models, Hugging Face, Ollama - Типи моделей: Phi-3, Phi-3.5, Phi-4 (текстові, зору,…
PhiCookBook ایک جامع کوک بک ریپوزٹری ہے جس میں Microsoft کے Phi خاندان کے چھوٹے زبان ماڈلز (SLMs) کے ساتھ کام کرنے کے لیے عملی مثالیں، ٹیوٹوریلز، اور دستاویزات شامل ہیں۔ یہ ریپوزٹری مختلف استعمال کے کیسز جیسے انفرنس، فائن-ٹیوننگ، کوانٹائزیشن، RAG نفاذات، اور مختلف پلیٹ فارمز اور فریم ورکس پر ملٹی موڈل ایپلی کیشنز کو ظاہر کرتی ہے۔ اہم ٹیکنالوجیز: - زبانیں: Python، C#/.NET، JavaScript/Node.js - فریم ورکس: ONNX Runtime، PyTorch، Transformers، MLX، OpenVINO، Semantic Kernel - پلیٹ فارمز:…
PhiCookBook là một kho lưu trữ sách hướng dẫn toàn diện chứa các ví dụ thực hành, bài hướng dẫn và tài liệu làm việc với dòng mô hình Ngôn Ngữ Nhỏ (SLMs) Phi của Microsoft. Kho lưu trữ minh họa nhiều trường hợp sử dụng khác nhau bao gồm suy luận, điều chỉnh tinh chỉnh, lượng tử hóa, triển khai RAG và các ứng dụng đa phương thức trên nhiều nền tảng và framework khác nhau. Công nghệ chính:…
A short guide for AI agents (and the humans who wire them up) that want to use tgrep as a fast search tool inside a repository. It complements the README, which documents every flag; this file covers the few things an agent has to get right. tgrep is ripgrep with a pre-built trigram index and an optional server. A search resolves in this order: 1. Server running for this tree: query it over TCP. Fastest. A file watcher keeps the…
A repository of skills, prompts, and MCP configurations for AI coding agents working with Azure SDKs and Microsoft AI Foundry services. Azure SDKs and Foundry APIs change constantly. Never work with stale knowledge. Before implementing anything with Azure/Foundry SDKs: If you skip this step and use outdated patterns, you will produce broken code. These principles reduce common LLM coding mistakes. Apply them to every task. Don't assume. Don't hide confusion. Surface tradeoffs. Minimum code that solves the problem. Nothing speculative.
Generate AGENTS.md files for pertinent repository folders — only where missing. Provides coding agents with build commands, testing instructions, code style, project structure, and boundaries.
This folder contains a test harness for evaluating AI-generated code against acceptance criteria for skills. What we're testing: Skills in .github/skills/ that provide domain knowledge for Azure SDKs. How it works: Run pnpm harness --list from the tests/ directory to see all skills with criteria. Location: tests/scenarios/<skill-name>/acceptance-criteria.md Source materials (in order of priority): Critical: Document import distinctions carefully. Many Azure SDKs have models in different locations (e.g., azure.ai.agents.models vs azure.ai.projects.models). Location: tests/scenarios/<skill-name>/scenarios.yaml Scenario design principles: Success criteria:
Olive is a Python toolkit for AI model optimization on ONNX Runtime. The main package is in olive/, tests are in test/, and the CLI entry point is olive. Olive users should install olive-ai with pip install olive-ai, and follow README.md. Use the smallest relevant check for your change: If you need to run tests, install the appropriate dependencies from test/requirements-test.txt or one of the platform-specific files in test/.
Intelligent Terminal is a Windows Terminal fork that adds first-class AI agent workflows. The inherited Windows Terminal build, architecture, and C++ conventions are documented in .github/copilot-instructions.md; this file contains only the fork-specific context. - WTA (tools/wta/) is the Rust orchestrator. - ACP means Agent Client Protocol. wta-master lazily owns a pool of agent CLI processes keyed by agent identity, execution source, and command; helpers using the same key share one process and multiplex sessions through it. - WT Protocol is…
These rules apply to tools/wta/. The repo-level AGENTS.md is the canonical source for build, deployment, supported providers, runtime paths, and log names. Do not duplicate those facts here. WTA is a Rust binary with three launch modes: - Master (--master <pipe>): lazily owns a pool of agent CLI subprocesses and their ACP stdio connections, accepts helper named-pipe connections, and routes each ACP SessionId to its owning helper. Helpers with the same master-derived agent key share a process. Implementation: src/master/mod.rs. -…
Just is a Microsoft build task organization library for JavaScript/TypeScript projects. It provides a task runner (just-task) and preset build scripts (just-scripts) for common workflows (TypeScript, Webpack, Jest, ESLint, etc.). DO NOT run jest directly. You MUST run yarn test (which wraps jest and takes all the same args) from the specific package to ensure the correct environment is set up. Built on Undertaker (Gulp's task orchestrator). Key modules in packages/just-task/src/: Task functions can: return a Promise, call a done…
The extension integrates ESLint into VS Code by implementing a ESLint server using the Language Server Protocol (LSP). Key goals: * Defer activation until needed (lazy activation based on opened documents / settings). * Support multiple workspace folders & dynamic working directory inference. * Abstract ESLint version / API differences (CLIEngine vs ESLint class / flat config) transparently. * Provide rich code actions (fix all, rule disable, suggestions) and formatting integration. * Handle notebooks and non-file schemes where possible. Settings…
This is the server implementation of the ESLint extension The VS Code API uses URIs to identify resource. ESLint in contrast uses file system paths. The conversion from URIs into file system paths happens in the function getFileSystemPath in the server\src\paths.ts file
MicrosoftAGENTS.md