RecIS Documentation

Note

语言选择

中文文档: RecIS 文档

RecIS (Recommendation Intelligence System) is a unified architecture deep learning framework designed specifically for ultra-large-scale sparse + multimodal + large language models, built on the PyTorch open-source ecosystem.

Quick Navigation

Chinese Documentation	English Documentation
项目介绍	Project Introduction
安装指南	Installation Guide
快速开始	Quick Start
API Documentation	API Documentation
示例教程	Example Tutorials
Tools Documentation	Tools Documentation
常见问题	Frequently Asked Questions

Quick Start

Installation: Follow the Installation Guide guide to set up RecIS
Quick Tutorial: Try the Quick Start tutorial to build your first model
Examples: Browse Example Tutorials for detailed use cases
API Reference: Check API Documentation for complete API documentation

Core Features

Ultra-large Scale Support: Unified architecture for ultra-large-scale sparse models
High Performance: Built on PyTorch open-source ecosystem with sparse model sharding optimization
Ease to Use: Provides clean API design with native computation optimization techniques
Compatibility: Fully compatible with existing PyTorch ecosystem

Table of Contents

Contents: