[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-RedisAI--redis-inference-optimization":3,"tool-RedisAI--redis-inference-optimization":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":75,"owner_avatar_url":76,"owner_bio":77,"owner_company":78,"owner_location":78,"owner_email":78,"owner_twitter":79,"owner_website":80,"owner_url":81,"languages":82,"stars":107,"forks":108,"last_commit_at":109,"license":110,"difficulty_score":10,"env_os":111,"env_gpu":112,"env_ram":113,"env_deps":114,"category_tags":121,"github_topics":122,"view_count":23,"oss_zip_url":78,"oss_zip_packed_at":78,"status":16,"created_at":128,"updated_at":129,"faqs":130,"releases":159},1982,"RedisAI\u002Fredis-inference-optimization","redis-inference-optimization","A Redis module for serving tensors and executing deep learning graphs","Redis-inference-optimization 是一个 Redis 模块，用于直接在 Redis 中加载和运行深度学习模型，支持 TensorFlow、PyTorch、ONNXRuntime 等主流框架。它让模型推理不再依赖外部服务，而是与数据存储在同一系统内完成，显著降低延迟、提升吞吐量，特别适合需要高频、低延迟推理的场景。通过利用 Redis 高效的数据本地化机制，它避免了模型与数据在不同服务间频繁传输的开销，简化了生产环境的部署流程。该模块适合有机器学习部署经验的开发者或数据工程师，尤其是那些已在使用 Redis 作为缓存或消息系统、希望将 AI 推理无缝集成进现有架构的团队。需要注意的是，该项目已于 2025 年停止维护，原名 RedisAI，官方建议转向 Redis 官方最新 AI 解决方案。如需使用，建议参考其 1.2.7 版本的 Docker 镜像或源码构建，同时注意模型后端版本的兼容性。","> [!CAUTION]\n> **Redis-inference-optimization is no longer actively maintained or supported.**\n>\n> We are grateful to the redis-inference-optimization community for their interest and support.\n> Previously, redis-inference-optimization was named RedisAI, but was renamed in Jan 2025 to reduce confusion around Redis' other AI offerings. To learn more about Redis' current AI offerings, visit [the Redis website](https:\u002F\u002Fredis.io\u002Fredis-for-ai).\n\n# Redis-inference-optimization\nRedis-inference-optimization is a Redis module for executing Deep Learning\u002FMachine Learning models and managing their data. Its purpose is being a \"workhorse\" for model serving, by providing out-of-the-box support for popular DL\u002FML frameworks and unparalleled performance. **Redis-inference-optimization both maximizes computation throughput and reduces latency by adhering to the principle of data locality**, as well as simplifies the deployment and serving of graphs by leveraging on Redis' production-proven infrastructure.\n\n# Quickstart\nRedis-inference-optimization is a Redis module. To run it you'll need a Redis server (v6.0.0 or greater), the module's shared library, and its dependencies.\n\nThe following sections describe how to get started with redis-inference-optimization.\n\n## Docker\nThe quickest way to try redis-inference-optimization is by launching its official Docker container images.\n### On a CPU only machine\n```\ndocker run -p 6379:6379 redislabs\u002Fredisai:1.2.7-cpu-bionic\n```\n\n### On a GPU machine\nFor GPU support you will need a machine you'll need a machine that has Nvidia driver (CUDA 11.3 and cuDNN 8.1), nvidia-container-toolkit and Docker 19.03+ installed. For detailed information, checkout [nvidia-docker documentation](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fnvidia-docker)\n\n```\ndocker run -p 6379:6379 --gpus all -it --rm redislabs\u002Fredisai:1.2.7-gpu-bionic\n```\n\n\n## Building\nYou can compile and build the module from its source code. \n\n### Prerequisites\n* Packages: git, python3, make, wget, g++\u002Fclang, & unzip\n* CMake 3.0 or higher needs to be installed.\n* CUDA 11.3 and cuDNN 8.1 or higher needs to be installed if GPU support is required.\n* Redis v6.0.0 or greater.\n\n### Get the Source Code\nYou can obtain the module's source code by cloning the project's repository using git like so:\n\n```sh\ngit clone --recursive https:\u002F\u002Fgithub.com\u002FRedisAI\u002Fredis-inference-optimization\n```\n\nSwitch to the project's directory with:\n\n```sh\ncd redis-inference-optimization\n```\n\n### Building the Dependencies\nUse the following script to download and build the libraries of the various redis-inference-optimization backends (TensorFlow, PyTorch, ONNXRuntime) for CPU only:\n\n```sh\nbash get_deps.sh\n```\n\nAlternatively, you can run the following to fetch the backends with GPU support.\n\n```sh\nbash get_deps.sh gpu\n```\n\n### Building the Module\nOnce the dependencies have been built, you can build the redis-inference-optimization module with:\n\n```sh\nmake -C opt clean ALL=1\nmake -C opt\n```\n\nAlternatively, run the following to build redis-inference-optimization with GPU support:\n\n```sh\nmake -C opt clean ALL=1\nmake -C opt GPU=1\n```\n\n### Backend Dependancy\n\nRedis-inference-optimization currently supports PyTorch (libtorch), Tensorflow (libtensorflow), TensorFlow Lite, and ONNXRuntime as backends. This section shows the version map between redis-inference-optimization and supported backends. This extremely important since the serialization mechanism of one version might not match with another. For making sure your model will work with a given redis-inference-optimization version, check with the backend documentation about incompatible features between the version of your backend and the version redis-inference-optimization is built with.\n\n\n| redis-inference-optimization | PyTorch  | TensorFlow | TFLite | ONNXRuntime |\n|:--------|:--------:|:----------:|:------:|:-----------:|\n| 1.0.3   |  1.5.0   |   1.15.0   | 2.0.0  |    1.2.0    |\n| 1.2.7   |  1.11.0  |   2.8.0    | 2.0.0  |   1.11.1    |\n| master  |  1.11.0  |   2.8.0    | 2.0.0  |   1.11.1    |\n\nNote: Keras and TensorFlow 2.x are supported through graph freezing. \n\n## Loading the Module\nTo load the module upon starting the Redis server, simply use the `--loadmodule` command line switch, the `loadmodule` configuration directive or the [Redis `MODULE LOAD` command](https:\u002F\u002Fredis.io\u002Fcommands\u002Fmodule-load) with the path to module's library.\n\nFor example, to load the module from the project's path with a server command line switch use the following:\n\n```sh\nredis-server --loadmodule .\u002Finstall-cpu\u002Fredis-inference-optimization.so\n```\n\n### Give it a try\n\nOnce loaded, you can interact with redis-inference-optimization using redis-cli. \n\n### Client libraries\nSome languages already have client libraries that provide support for redis-inference-optimization's commands. The following table lists the known ones:\n\n| Project            | Language              | License      | Author                                           | URL                                                         |\n| -------            | --------              | -------      | ------                                           | ---                                                         |\n| JredisAI           | Java                  | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FRedisAI\u002FJRedisAI)               |\n| redisAI-py         | Python                | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FredisAI\u002FredisAI-py)             |\n| redisAI-go         | Go                    | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FRedisAI\u002FredisAI-go)             |\n| redisAI-js         | Typescript\u002FJavascript | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FredisAI\u002FredisAI-js)             |\n| redis-modules-sdk  | TypeScript            | BSD-3-Clause | [Dani Tseitlin](https:\u002F\u002Fgithub.com\u002Fdanitseitlin) | [Github](https:\u002F\u002Fgithub.com\u002Fdanitseitlin\u002Fredis-modules-sdk) |\n| redis-modules-java | Java                  | Apache-2.0   | [dengliming](https:\u002F\u002Fgithub.com\u002Fdengliming)      | [Github](https:\u002F\u002Fgithub.com\u002Fdengliming\u002Fredis-modules-java)  |\n| smartredis         | C++                   | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n| smartredis         | C                     | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n| smartredis         | Fortran               | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n| smartredis         | Python                | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n\n## License\nRedis-inference-optimization is licensed under your choice of the Redis Source Available License 2.0 (RSALv2) or the Server Side Public License v1 (SSPLv1).\n","> [!CAUTION]\n> **Redis推理优化现已不再积极维护或支持。**\n>\n> 我们感谢Redis推理优化社区对它的关注与支持。\n> 此前，Redis推理优化名为RedisAI，但为减少与Redis其他AI产品的混淆，已于2025年1月更名。如需了解Redis当前的AI产品，请访问[Redis官网](https:\u002F\u002Fredis.io\u002Fredis-for-ai)。\n\n# Redis推理优化\nRedis推理优化是一个用于执行深度学习\u002F机器学习模型并管理其数据的Redis模块。它的目标是成为模型服务的“主力”，通过开箱即用的支持热门DL\u002FML框架以及无与伦比的性能来实现这一目标。**Redis推理优化遵循数据局部性原则，既能最大化计算吞吐量，又能降低延迟**，同时借助Redis久经考验的基础设施，简化了图的部署与服务流程。\n\n# 快速入门\nRedis推理优化是一个Redis模块。要运行它，您需要一个Redis服务器（版本6.0.0或更高）、该模块的共享库及其依赖项。\n\n以下各节将介绍如何快速开始使用Redis推理优化。\n\n## Docker\n尝试Redis推理优化最便捷的方式是启动其官方Docker容器镜像。\n### 仅在CPU机器上\n```\ndocker run -p 6379:6379 redislabs\u002Fredisai:1.2.7-cpu-bionic\n```\n\n### 在GPU机器上\n若要支持GPU，您需要一台已安装Nvidia驱动（CUDA 11.3和cuDNN 8.1）、nvidia-container-toolkit以及Docker 19.03+的机器。有关详细信息，请查看[Nvidia Docker文档](https:\u002F\u002Fgithub.com\u002FNVIDIA\u002Fnvidia-docker)。\n\n```\ndocker run -p 6379:6379 --gpus all -it --rm redislabs\u002Fredisai:1.2.7-gpu-bionic\n```\n\n\n## 构建\n您可以从源代码编译并构建该模块。\n\n### 前提条件\n* 软件包：git、python3、make、wget、g++\u002Fclang及unzip\n* 需安装CMake 3.0或更高版本。\n* 若需GPU支持，需安装CUDA 11.3和cuDNN 8.1或更高版本。\n* Redis版本6.0.0或更高。\n\n### 获取源代码\n您可以通过git克隆项目仓库获取模块的源代码，如下所示：\n\n```sh\ngit clone --recursive https:\u002F\u002Fgithub.com\u002FRedisAI\u002Fredis-inference-optimization\n```\n\n切换到项目目录：\n\n```sh\ncd redis-inference-optimization\n```\n\n### 构建依赖项\n使用以下脚本下载并构建各种Redis推理优化后端（TensorFlow、PyTorch、ONNXRuntime）的库，仅限CPU：\n\n```sh\nbash get_deps.sh\n```\n\n或者，您也可以运行以下命令以获取支持GPU的后端。\n\n```sh\nbash get_deps.sh gpu\n```\n\n### 构建模块\n依赖项构建完成后，您可以使用以下命令构建Redis推理优化模块：\n\n```sh\nmake -C opt clean ALL=1\nmake -C opt\n```\n\n或者，运行以下命令以构建支持GPU的Redis推理优化模块：\n\n```sh\nmake -C opt clean ALL=1\nmake -C opt GPU=1\n```\n\n### 后端依赖\n\nRedis推理优化目前支持PyTorch（libtorch）、TensorFlow（libtensorflow）、TensorFlow Lite和ONNXRuntime作为后端。本节展示了Redis推理优化与支持后端之间的版本对应关系。这一点极为重要，因为不同版本的序列化机制可能不兼容。为确保您的模型能与特定版本的Redis推理优化配合使用，请查阅后端文档，确认您的后端版本与Redis推理优化构建版本之间是否存在不兼容功能。\n\n\n| Redis推理优化 | PyTorch | TensorFlow | TFLite | ONNXRuntime |\n|:--------|:--------:|:----------:|:------:|:-----------:|\n| 1.0.3   |  1.5.0   |   1.15.0   | 2.0.0  |    1.2.0    |\n| 1.2.7   |  1.11.0  |   2.8.0    | 2.0.0  |   1.11.1    |\n| master  |  1.11.0  |   2.8.0    | 2.0.0  |   1.11.1    |\n\n注：Keras和TensorFlow 2.x可通过图冻结方式支持。\n\n## 加载模块\n要在启动Redis服务器时加载模块，只需使用`--loadmodule`命令行开关、`loadmodule`配置指令或[Redis `MODULE LOAD`命令](https:\u002F\u002Fredis.io\u002Fcommands\u002Fmodule-load)，并提供模块库的路径即可。\n\n例如，使用服务器命令行开关从项目路径加载模块，可执行以下命令：\n\n```sh\nredis-server --loadmodule .\u002Finstall-cpu\u002Fredis-inference-optimization.so\n```\n\n### 试试看\n\n加载完成后，您可使用redis-cli与Redis推理优化进行交互。\n\n### 客户端库\n一些语言已有支持Redis推理优化命令的客户端库。下表列出了已知的几个：\n\n| 项目            | 语言              | 许可证      | 作者                                           | URL                                                         |\n| -------            | --------              | -------      | ------                                           | ---                                                         |\n| JredisAI           | Java                  | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FRedisAI\u002FJRedisAI)               |\n| redisAI-py         | Python                | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FredisAI\u002FredisAI-py)             |\n| redisAI-go         | Go                    | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FredisAI\u002FredisAI-go)             |\n| redisAI-js         | Typescript\u002FJavascript | BSD-3        | [Redis](https:\u002F\u002Fredis.io)                        | [Github](https:\u002F\u002Fgithub.com\u002FredisAI\u002FredisAI-js)             |\n| redis-modules-sdk  | TypeScript            | BSD-3-Clause | [Dani Tseitlin](https:\u002F\u002Fgithub.com\u002Fdanitseitlin) | [Github](https:\u002F\u002Fgithub.com\u002Fdanitseitlin\u002Fredis-modules-sdk) |\n| redis-modules-java | Java                  | Apache-2.0   | [dengliming](https:\u002F\u002Fgithub.com\u002Fdengliming)      | [Github](https:\u002F\u002Fgithub.com\u002Fdengliming\u002Fredis-modules-java)  |\n| smartredis         | C++                   | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n| smartredis         | C                     | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n| smartredis         | Fortran               | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n| smartredis         | Python                | BSD-2-Clause | [Cray Labs](https:\u002F\u002Fgithub.com\u002FCrayLabs)         | [Github](https:\u002F\u002Fgithub.com\u002FCrayLabs\u002FSmartRedis)            |\n\n## 许可证\nRedis推理优化采用您选择的Redis源码可用许可证2.0（RSALv2）或服务器端公共许可证v1（SSPLv1）进行许可。","# Redis-inference-optimization 快速上手指南\n\n> ⚠️ 注意：Redis-inference-optimization 已停止维护。原名为 RedisAI，2025 年 1 月更名。建议移步 [Redis 官方 AI 产品页](https:\u002F\u002Fredis.io\u002Fredis-for-ai) 获取最新支持。\n\n---\n\n## 环境准备\n\n- **系统要求**：Linux（推荐 Ubuntu 18.04\u002F20.04）\n- **Redis 版本**：v6.0.0 或更高\n- **CPU 模式依赖**：git、python3、make、wget、g++\u002Fclang、unzip、CMake 3.0+\n- **GPU 模式依赖**：\n  - NVIDIA 驱动 + CUDA 11.3 + cuDNN 8.1+\n  - nvidia-container-toolkit\n  - Docker 19.03+\n\n> 推荐使用 Docker 快速部署，避免编译复杂依赖。\n\n---\n\n## 安装步骤\n\n### 方案一：Docker（推荐）\n\n#### CPU 环境\n```bash\ndocker run -p 6379:6379 redislabs\u002Fredisai:1.2.7-cpu-bionic\n```\n\n#### GPU 环境（需 NVIDIA 支持）\n```bash\ndocker run -p 6379:6379 --gpus all -it --rm redislabs\u002Fredisai:1.2.7-gpu-bionic\n```\n\n> 国内用户可使用阿里云镜像加速：\n> ```bash\n> docker run -p 6379:6379 redislabs\u002Fredisai:1.2.7-cpu-bionic --registry-mirror=https:\u002F\u002F\u003Cyour-mirror>.mirror.aliyuncs.com\n> ```\n\n### 方案二：源码编译（高级用户）\n\n```bash\ngit clone --recursive https:\u002F\u002Fgithub.com\u002FRedisAI\u002Fredis-inference-optimization\ncd redis-inference-optimization\n\n# CPU 版本依赖\nbash get_deps.sh\n\n# GPU 版本依赖（需提前安装 CUDA\u002FcuDNN）\nbash get_deps.sh gpu\n\n# 编译模块\nmake -C opt clean ALL=1\nmake -C opt  # CPU\n# 或\nmake -C opt GPU=1  # GPU\n```\n\n加载模块：\n```bash\nredis-server --loadmodule .\u002Finstall-cpu\u002Fredis-inference-optimization.so\n```\n\n---\n\n## 基本使用\n\n启动 Redis 后，使用 `redis-cli` 测试模型推理：\n\n```bash\nredis-cli\n```\n\n在 redis-cli 中执行：\n\n```bash\nAI.TENSORSET mytensor FLOAT 2 2 VALUES 1 2 3 4\nAI.MODELRUN mymodel INPUTS mytensor OUTPUTS output\nAI.TENSORGET output VALUES\n```\n\n> 模型需预先使用 `AI.MODELSET` 加载（如 TensorFlow\u002FPyTorch 模型），详见官方文档。\n\n支持客户端库：Python（redisAI-py）、Java（JredisAI）、Go（redisAI-go）等，可通过 GitHub 安装对应 SDK。","一家电商公司正在实时推荐系统中部署一个基于PyTorch的用户点击预测模型，该模型需在毫秒级响应时间内为每位访客生成个性化商品推荐。系统每日处理数亿次请求，对延迟和吞吐量极为敏感。\n\n### 没有 redis-inference-optimization 时\n- 模型部署在独立的Python服务中，每次推理需通过HTTP调用，网络延迟高达50–100ms，严重影响推荐响应速度。\n- 模型权重和输入张量需在内存与磁盘间频繁序列化\u002F反序列化，增加CPU负担，导致服务资源利用率低下。\n- 多个微服务各自加载相同模型副本，内存占用高达数GB，运维成本高且易出现版本不一致。\n- 模型更新需停机重启服务，无法实现热加载，影响线上稳定性。\n- 缺乏统一的模型管理接口，监控、日志和负载均衡需额外开发，开发周期延长两周以上。\n\n### 使用 redis-inference-optimization 后\n- 模型直接加载进Redis，推理在内存中完成，端到端延迟降至5ms以内，推荐响应速度提升90%。\n- 张量数据无需序列化传输，直接在Redis内部执行计算，CPU开销降低60%，服务器资源可节省40%。\n- 多个推荐服务共享同一模型实例，内存占用从8GB降至1.2GB，部署一致性100%保障。\n- 支持通过Redis命令动态加载新模型版本，无需重启服务，上线时间从小时级缩短至秒级。\n- 内置监控指标和高可用架构，可无缝接入现有Redis监控体系，运维复杂度下降70%。\n\nredis-inference-optimization 将模型推理从独立服务转变为内存级原生能力，让实时推荐系统在性能、成本与稳定性上实现质的飞跃。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002FRedisAI_redis-inference-optimization_39ccfc8c.png","RedisAI","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002FRedisAI_a3da518b.png","",null,"Redisinc","https:\u002F\u002Fredisai.io","https:\u002F\u002Fgithub.com\u002FRedisAI",[83,87,91,95,99,103],{"name":84,"color":85,"percentage":86},"C","#555555",62.4,{"name":88,"color":89,"percentage":90},"Python","#3572A5",27.6,{"name":92,"color":93,"percentage":94},"C++","#f34b7d",5.7,{"name":96,"color":97,"percentage":98},"Shell","#89e051",2.1,{"name":100,"color":101,"percentage":102},"CMake","#DA3434",1.2,{"name":104,"color":105,"percentage":106},"Makefile","#427819",0.9,841,108,"2026-03-19T04:38:41","NOASSERTION","Linux","需要 NVIDIA GPU，CUDA 11.3，cuDNN 8.1，显存未明确说明","未说明",{"notes":115,"python":116,"dependencies":117},"该工具已停止维护，建议使用 Redis 官方当前 AI 方案；构建时需安装 git、make、g++\u002Fclang、CMake 3.0+；支持 CPU 和 GPU 模式，GPU 模式需安装 nvidia-container-toolkit 和 Docker 19.03+；模型后端版本需严格匹配，否则可能导致序列化不兼容","3+",[118,119,120],"libtorch","libtensorflow","onnxruntime",[13],[123,124,125,120,126,127],"redisai","pytorch","tensorflow","serving-tensors","machine-learning","2026-03-27T02:49:30.150509","2026-04-06T05:16:53.279158",[131,136,141,146,151,155],{"id":132,"question_zh":133,"answer_zh":134,"source_url":135},8942,"RedisAI 的命令格式在新版本中如何统一？","新版本已统一命令格式为：AI.MODELSET、AI.MODELRUN、AI.SCRIPTSET、AI.SCRIPTGET、AI.SCRIPTRUN、AI.TENSORSET、AI.TENSORGET。不再使用 AI.SET 和 AI.RUN，而是明确区分模型、脚本和张量操作。例如：AI.MODELSET model TF CPU BLOB \u003C model.pb，AI.MODELRUN model INPUTS x OUTPUTS y。","https:\u002F\u002Fgithub.com\u002FRedisAI\u002Fredis-inference-optimization\u002Fissues\u002F49",{"id":137,"question_zh":138,"answer_zh":139,"source_url":140},8943,"在 Redis 集群模式下 AI.SCRIPTRUN 报错如何解决？","该问题已在 v1.0 和 v1.2 版本中修复。请升级到 RedisAI 的最新版本（如 edge 镜像），并确保使用包含 no-cluster 标志的命令实现。可使用 docker pull redisai\u002Fredisai:edge 获取修复版本，日志中应显示 git_sha=b1494aa0386b42527a941356840c70e1bede963d。","https:\u002F\u002Fgithub.com\u002FRedisAI\u002Fredis-inference-optimization\u002Fissues\u002F437",{"id":142,"question_zh":143,"answer_zh":144,"source_url":145},8944,"AI.MODELSET 报错 'Invalid GraphDef' 是什么原因？","这是因为下载的模型文件是 Git LFS 的链接而非实际二进制文件。必须执行 git-lfs pull 命令来拉取真实的模型文件（如 graph.pb），再通过 AI.MODELSET 命令加载。直接使用 git pull 会得到无效的文本链接，导致解析失败。","https:\u002F\u002Fgithub.com\u002FRedisAI\u002Fredis-inference-optimization\u002Fissues\u002F136",{"id":147,"question_zh":148,"answer_zh":149,"source_url":150},8945,"YOLO 模型在 RedisAI 上推理耗时 5 秒以上，如何优化？","该性能问题已在后续版本中显著改善。请升级到最新版 RedisAI 并使用 GPU 容器（如 redisai\u002Fredisai:edge），在 RTX 2070 上推理时间可降至 0.025–0.08 秒。确保使用 GPU 模式运行，并避免在 CPU 上执行推理任务。","https:\u002F\u002Fgithub.com\u002FRedisAI\u002Fredis-inference-optimization\u002Fissues\u002F119",{"id":152,"question_zh":153,"answer_zh":154,"source_url":140},8946,"如何在 Redis 集群中正确使用 AI.TENSORSET 和 AI.TENSORGET？","AI.TENSORSET 和 AI.TENSORGET 命令在 Redis 集群模式下（使用 -c 参数）是支持的。确保 RedisAI 版本为 v1.0 或更高，且命令语法正确，例如：redis-cli -c AI.TENSORSET key FLOAT 1 28 28 BLOB \u003C data.raw。这些命令会自动根据 key 的 hash slot 分发到正确节点。",{"id":156,"question_zh":157,"answer_zh":158,"source_url":140},8947,"如何在不使用 Docker 的情况下部署 RedisAI 到 Slurm 集群？","可从源码编译 RedisAI 模块。克隆 RedisAI 仓库（如 https:\u002F\u002Fgithub.com\u002FRedisAI\u002FRedisAI），按文档编译生成 redisai.so，然后在 Redis 配置文件中通过 loadmodule 指令加载该模块。确保系统已安装 CUDA、cuDNN 和 PyTorch\u002FTensorFlow 开发依赖。",[160,165,170,175,180,185,190,195,200,205,210,215,220,225,230,235,239],{"id":161,"version":162,"summary_zh":163,"released_at":164},106379,"v1.2.7","This is a maintenance release for RedisAI 1.2\r\nUpdate urgency: Medium\r\n\r\nHeadlines:\r\nThis release improves overall stability and provides fixes for issues found after the previous release.\r\n\r\n## Details\r\n\r\n### Minor enhancements:\r\n- Add AI.CONFIG GET sub-command (#918, #924)\r\n- Backends update - TF 2.8, PyTorch 1.11, ONNXRuntime 1.11 (#914, #915, #917)\r\n- Enable saving model\u002Fscript run statistics when execution occurs from low-level API (by using RedisGears integration in particular), and retrieving the statistics with AI.INFO command (#904)\r\n- Restore support for MacOS build scripts (#897)\r\n- Removed support for Linux Xenial (#893) \r\n\r\n### Bugfixes:\r\n- Fix a synchronization issue regarding update the number of background threads, that caused (rarely) a crash upon executing models in ONNX (#923)","2022-06-30T12:21:29",{"id":166,"version":167,"summary_zh":168,"released_at":169},106380,"v1.2.5","\r\n## Headlines\r\n- (#832) Strings tensor support (TF and ONNXRUNTIME only).\r\n- (#847) (#806) Backends update - TF 2.6, PyTorch 1.9, ONNXRuntime 1.9\r\n- (#827) ONNXRuntime memory is now managed by Redis (Cloud readiness)\r\n\r\n## Bugfixes\r\n- (#829) Remove deprecation warnings from deprecated commands on Redis logs.\r\n- (#852) fixed invalid delete of outputs after execution error in TF \r\n\r\n\r\n","2021-11-01T09:53:06",{"id":171,"version":172,"summary_zh":173,"released_at":174},106381,"v1.2.4","This is the fourth release candidate for RedisAI v1.2!\r\n\r\n## Headlines:\r\n#482 DAG performance enhancement: Auto-batching support for MODELRUN commands inside a DAG Added DAG general timeout.\r\n#489 Execute Redis commands in TorchScript. This capability enables, but not limited, to convert data residing in Redis (or modules) data structures into tensors to be fed to the model as well as store script results in Redis data structures. Scripts can be run in DAG's.\r\n#787 Allow executing models from within TorchScript.\r\n#775 Support for `BOOL` tensors.\r\n#511, #537, #547, #556 Allow async model, script and DAG run via low-level API. This capability will allow other modules to call RedisAI. For example a module that hold time series data, can call RedisAI directly for anomaly detection.\r\n#723 #682 #680, #792  - New API commands `MODELSTORE` `MODELEXECUTE`, `SCRIPTSTORE`, `SCRIPTEXECUTE` `DAGEXECUTE` replacing and deprecating `MODELSET` `MODELRUN`, `SCRIPTSET`, `SCRIPTRUN` `DAGRUN` for better enterprise cluster support. \r\n#779 #797  Allow ONNXRuntime backend execution timeout\r\n\r\n\r\n## Details:\r\n### Minor enhancements:\r\n#566 #619  TensorFlow 2.4, Pytorch 1.7, ONNXRuntime 1.7. Note: Current ONNXRuntime distribution is experimental under RedisAI build. Future version will return the original ONNXRuntime distribution.\r\n#499 Allow setting number of inter\u002Fintra op threads in Torch backend.\r\n#529 Expose model definitions for inputs and outputs in MODELGET command.\r\n#530 Expose Redis\u002FRedisAI main thread cpu usage.\r\n#540 Reuse memory in TENSORSET command.\r\n#580 Expose backend version on AI.INFO\r\n#581 Cache tensor length.\r\n#791 Allow `AI.MODELGET` and `AI.SCRIPTGET` without optional args\r\n\r\n### Bugfixes:\r\n#488 Handle short reads during RDB load.\r\n#538 Handle binary strings as tensor names.\r\n#462, #553 Handle memory leaks.\r\n#558 Erroneous replies for invalid `MODELGET` commands.\r\n#748 Model RDB encoding missing properties\r\n#754 AOF-Rewrite logic\r\n#793 Fix DAG reply for `AI.TENSORGET` op\r\n#794 #794 Fix `MODELGET`, `SCRIPTGET` , `_MODELSCAN` and `_SCRIPTSCAN` commands\r\n\r\nNotes:\r\nThis is not the GA version of 1.2. The version inside Redis will be 10204 or 1.2.4 in semantic versioning. Since the version of a module in Redis is numeric, we could not add an RC4 flag.","2021-07-27T09:06:46",{"id":176,"version":177,"summary_zh":178,"released_at":179},106382,"v1.2.3","This is the third release candidate for RedisAI v1.2!\r\n\r\n## Headlines:\r\n#482 DAG performance enhancement: Auto-batching support for MODELRUN commands inside a DAG Added DAG general timeout.\r\n#489 Execute Redis commands in TorchScript. This capability enables, but not limited, to convert data residing in Redis (or modules) data structures into tensors to be fed to the model as well as store script results in Redis data structures. Scripts can be run in DAG's.\r\n#511, #537, #547, #556 Allow async model, script and DAG run via low-level API. This capability will allow other modules to call RedisAI. For example a module that hold time series data, can call RedisAI directly for anomaly detection.\r\n#723 #682 #680 - New API commands `MODELSTORE` `MODELEXECUTE` `SCRIPTEXECUTE` `DAGEXECUTE` replacing and deprecating `MODELSET` `MODELRUN` `SCRIPTRUN` `DAGRUN` for better enterprise cluster support. Note: `SCRIPTEXECUTE` command structure will be changed in the following release candidates, as well as new command `SCRIPTSTORE` will be introduced.\r\n\r\n## Details:\r\n### Minor enhancements:\r\n#566 TensorFlow 2.4, Pytorch 1.7, ONNXRuntime 1.6.\r\n#499 Allow setting number of inter\u002Fintra op threads in Torch backend.\r\n#529 Expose model definitions for inputs and outputs in MODELGET command.\r\n#530 Expose Redis\u002FRedisAI main thread cpu usage.\r\n#540 Reuse memory in TENSORSET command.\r\n#580 Expose backend version on AI.INFO\r\n#581 Cache tensor length.\r\n\r\n### Bugfixes:\r\n#488 Handle short reads during RDB load.\r\n#538 Handle binary strings as tensor names.\r\n#462, #553 Handle memory leaks.\r\n#558 Erroneous replies for invalid MODELGET commands.\r\n#748 Model RDB encoding missing properties\r\n#754 AOF-Rewrite logic\r\n\r\nNotes:\r\nThis is not the GA version of 1.2. The version inside Redis will be 10203 or 1.2.3 in semantic versioning. Since the version of a module in Redis is numeric, we could not add an RC3 flag.","2021-06-03T14:29:01",{"id":181,"version":182,"summary_zh":183,"released_at":184},106383,"v1.2.1","This is the first release candidate for RedisAI v1.2!\r\n### Headlines:\r\n- #482 [`DAG`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun) performance enhancement: Auto-batching support for `MODELRUN` commands inside a DAG Added DAG general timeout.\r\n- #489 Execute Redis commands in TorchScript.  This capability enables, but not limited, to convert data residing in Redis (or modules) data structures into tensors to be fed to the model as well as store script results in Redis data structures.  Scripts can be run in DAG's.\r\n- #511, #537, #547, #556  Allow async model, script and DAG run via low-level API.  This capability will allow other modules to call RedisAI.  For example a module that hold time series data, can call RedisAI directly for anomaly detection.\r\n\r\n### Details:\r\n- Minor enhancements:\r\n  - #566 TensorFlow 2.4, Pytorch 1.7, ONNXRuntime 1.6.\r\n  - #499 Allow setting number of inter\u002Fintra op threads in Torch backend.\r\n  - #529 Expose model definitions for inputs and outputs in `MODELGET` command.\r\n  - #530 Expose Redis\u002FRedisAI main thread cpu usage.\r\n  - #540 Reuse memory in `TENSORSET` command.\r\n  - #580 Expose backend version on `AI.INFO`\r\n  - #581 Cache tensor length.\r\n\r\n- Bugfixes:\r\n  - #488 Handle short reads during RDB load.\r\n  - #538 Handle binary strings as tensor names.\r\n  - #462, #553 Handle memory leaks.\r\n  - #558  erroneous replies for invalid `MODELGET` commands.\r\n\r\nNotes:\r\nThis is not the GA version of 1.2. The version inside Redis will be 10201 or 1.2.1 in semantic versioning. Since the version of a module in Redis is numeric, we could not add an RC1 flag.","2021-02-16T13:24:56",{"id":186,"version":187,"summary_zh":188,"released_at":189},106384,"v1.0.2","This is a maintenance release for version 1.0.\r\nUpdate urgency: Medium\r\n\r\nHeadlines:\r\nThis release improves overall stability and provides fixes for issues found after the previous release.\r\n\r\nDetails:\r\n- Minor updates:\r\n  - #383 Enable [`AI.SCRIPTRUN`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aiscriptrun) inside [`AI.DAGRUN`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun)\r\n  - #395 Add support for variadic arguments to `AI.SCRIPTRUN`\r\n  - #400 Low level API additions for use in other modules (e.g. [RedisGears](redisgears.io))\r\n  - #396 Add relevant RedisAI config entries to the Redis INFO output.  Helpful for standard monitoring systems\r\n\r\n- Bug Fixes:\r\n  - #403 Atomic ref count\r\n  - #406 Avoid splitting outputs in batches when nbatches == 1\r\n  - #438 Fixed flagged as \"getkeys-api\" during the registration ( AI.DAGRUN, AI.DAGRUN_RO, AI.MODELRUN, AI.SCRIPTRUN )\r\n  - #449 Safely add to arrays\r\n  - #443 Segfault for `AI.DAGRUN` + `AI.TENSORSET`\r\n\r\n\r\n","2020-10-06T13:17:07",{"id":191,"version":192,"summary_zh":193,"released_at":194},106385,"v1.0.1","This is a maintenance release for version 1.0.\r\nUpdate urgency: Medium\r\n\r\n### Headlines:\r\nThis release improves overall stability and provides fixes for issues found after the previous release.\r\n\r\n### Details:\r\nBug Fixes:\r\n- 7f87f8534e70927d67f99b35dc6a97156761587f Allow inconsistent zero batch outputs.\r\n- #385,#382 [`AI.SCRIPTRUN`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aiscriptrun) results were being replicated twice.\r\n- #384 [`AI.MODELGET`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aimodelget) to return *inputs*, *outputs*, *batchsize*, and *minbatchsize*.\r\n- #412 Several memory leaks.","2020-07-02T11:23:46",{"id":196,"version":197,"summary_zh":198,"released_at":199},106386,"v1.0.0","This is the General Availability Release of RedisAI 1.0 (v1.0.0)!\r\n\r\n### Headlines:\r\n- Data locality decreases the end-to-end inference time.  RedisAI allows you to run your DL\u002FML models where the reference data for these models lives. RedisAI also allows you to persist intermediate states of a computation in-memory.\r\n- Support for multiple backends which enables you to compose computations across backends but also to decouple model creation and model serving.\r\n- Scale your AI serving infrastructure by scaling Redis.\r\n\r\n### Supported Backends:\r\n\r\n- TensorFlow Lite 2.0\r\n- TensorFlow 1.15.0\r\n- PyTorch 1.5\r\n- ONXXRuntime 1.2.0\r\n\r\n### Details: \r\n- New Features:\r\n  - #241, #270  **auto-batching** support. Requests from multiple clients can be automatically and transparently batched in a single request for increased CPU\u002FGPU efficiency during serving.\r\n  - #322 Add  [`AI.DAGRUN`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun). With the new  [`AI.DAGRUN`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun) (DAG as in direct acycilc graph) command we support the prescription of combinations of other AI.* commands in a single execution pass, where intermediate keys are never materialised to Redis.\r\n  - #334 Add  [`AI.DAGRUN_RO`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun_ro) command, a read-only variant of AI.DAGRUN  \r\n  - #338  [`AI.MODELSET`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aimodelset) Added the possibility to provide a model in chunks.\r\n  - #332 Standardized GET methods (TENSORGET,MODELGET,SCRIPTGET) replies (breaking change for clients)\r\n  - #331 Cache model blobs for faster serialization and thread-safety.\r\n\r\n- Minor Enhancements:\r\n  - #289 Memory access and leak fixes.\r\n  - #319 Documentation improvements.\r\n\r\n- Build Enhancements:\r\n  - #299 Coverage info.\r\n  - #273 Enable running valgrind\u002Fcallgrind on test platform\r\n- #277, #296 tests extension and refactoring per backend.\r\n\r\nNotes:\r\nThe version inside Redis will be 10000 or 1.0.0 in semantic versioning.","2020-05-12T18:32:04",{"id":201,"version":202,"summary_zh":203,"released_at":204},106387,"v0.99.0","This is a major milestone release for RedisAI v0.9.9.\r\n\r\nSupported Backends:\r\n- TensorFlow Lite 2.0\r\n- TensorFlow 1.15.0\r\n- PyTorch 1.5\r\n- ONXXRuntime 1.2.0\r\n\r\nNew Features:\r\n- #241, #270  **auto-batching** support. Requests from multiple clients can be automatically and transparently batched in a single request for increased CPU\u002FGPU efficiency during serving.\r\n- #322 Add  [`AI.DAGRUN`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun). With the new  [`AI.DAGRUN`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun) (DAG as in direct acycilc graph) command we support the prescription of combinations of other AI.* commands in a single execution pass, where intermediate keys are never materialised to Redis.\r\n- #334 Add  [`AI.DAGRUN_RO`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aidagrun_ro) command, a read-only variant of AI.DAGRUN  \r\n- #338  [`AI.MODELSET`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aimodelset) Added the possibility to provide a model in chunks.\r\n- #332 Standardized GET methods (TENSORGET,MODELGET,SCRIPTGET) replies (breaking change for clients)\r\n- #331 Cache model blobs for faster serialization and thread-safety.\r\n\r\nMinor Enhancements:\r\n- #289 Memory access and leak fixes.\r\n- #319 Documentation improvements.\r\n\r\nBuild Enhancements:\r\n- #299 Coverage info.\r\n- #273 Enable running valgrind\u002Fcallgrind on test platform\r\n- #277, #296 tests extension and refactoring per backend.\r\n\r\nNotes:\r\nThe version inside Redis will be 9900 or 0.99.0 in semantic versioning. Since the version of a module in Redis is numeric, we use 0.99 to resemble that it's almost 1.0","2020-05-04T17:43:26",{"id":206,"version":207,"summary_zh":208,"released_at":209},106388,"v0.9.0","This is a major milestone release for RedisAI v0.9.\r\n\r\nSupported Backends:\r\n- TensorFlow Lite 2.0\r\n- TensorFlow 1.14.0\r\n- PyTorch 1.3.1\r\n- ONXXRuntime 1.0.0\r\n\r\nNew Features:\r\n- #255 AOF Support and replication without rerunning the inference on the slave.\r\n- #259 [`AI.INFO`](https:\u002F\u002Foss.redislabs.com\u002Fredisai\u002Fcommands\u002F#aiinfo) command with an initial set of statistics.\r\n\r\nBuild Enhancements:\r\n- #264 Several CI infrastructure enhancements.\r\n- #230 CI for GPU.\r\n- #265 MAC build.\r\n- #191 ARM builds.","2020-02-06T12:13:40",{"id":211,"version":212,"summary_zh":213,"released_at":214},106389,"v0.3.1","Release fixes #148.  Docker container failed loading libgomp.","2019-06-22T16:56:01",{"id":216,"version":217,"summary_zh":218,"released_at":219},106390,"v0.3.0","- Added functionality\r\n  - #141 PyTorch 1.1 support\r\n  - #138 ONNXRuntime support.  This comes with an example Jupyter notebook #139\r\n- Minor Bugfixes:\r\n  - #127 Crash on MODELRUN if INPUTS\u002FOUTPUTS names are not present in the TF graph\r\n  - #125 Segfault on RDBSave if buffer is freed\r\n  - #108 Memory leak in queuePush\r\n- Notes for GPU Mode:\r\n  - Tensorflow (1.12.0) and PyTorch (1.1.0) require CUDA 9.0. \r\n  - ONNXRuntime (0.4.0) requires CUDA 9.1\r\n","2019-06-20T08:49:54",{"id":221,"version":222,"summary_zh":223,"released_at":224},106391,"v0.2.1","This is a hotfix for v0.2.0 that fixes a silent bug when adding params to the run context, which would manifest when inputs or outputs exceed PARAM_INITIAL_SIZE.","2019-04-06T16:22:20",{"id":226,"version":227,"summary_zh":228,"released_at":229},106392,"v0.2.0","This is the preview release of RedisAI.  Includes support for Tensorflow, Pytorch and TorchScript","2019-03-27T11:32:13",{"id":231,"version":232,"summary_zh":233,"released_at":234},106393,"v0.1.0","This is the first beta release of RedisAI. Please check the docs on redisai.io\r\n\r\n","2019-03-05T18:39:15",{"id":236,"version":237,"summary_zh":78,"released_at":238},106394,"0.1.0-beta1","2019-03-03T12:31:53",{"id":240,"version":241,"summary_zh":78,"released_at":242},106395,"0.1.0-alpha1","2019-01-20T11:54:02"]