[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-uxlfoundation--oneDAL":3,"tool-uxlfoundation--oneDAL":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},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,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},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",143909,2,"2026-04-07T11:33:18",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},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。",107888,"2026-04-06T11:32:50",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":10,"last_commit_at":59,"category_tags":60,"status":17},4487,"LLMs-from-scratch","rasbt\u002FLLMs-from-scratch","LLMs-from-scratch 是一个基于 PyTorch 的开源教育项目，旨在引导用户从零开始一步步构建一个类似 ChatGPT 的大型语言模型（LLM）。它不仅是同名技术著作的官方代码库，更提供了一套完整的实践方案，涵盖模型开发、预训练及微调的全过程。\n\n该项目主要解决了大模型领域“黑盒化”的学习痛点。许多开发者虽能调用现成模型，却难以深入理解其内部架构与训练机制。通过亲手编写每一行核心代码，用户能够透彻掌握 Transformer 架构、注意力机制等关键原理，从而真正理解大模型是如何“思考”的。此外，项目还包含了加载大型预训练权重进行微调的代码，帮助用户将理论知识延伸至实际应用。\n\nLLMs-from-scratch 特别适合希望深入底层原理的 AI 开发者、研究人员以及计算机专业的学生。对于不满足于仅使用 API，而是渴望探究模型构建细节的技术人员而言，这是极佳的学习资源。其独特的技术亮点在于“循序渐进”的教学设计：将复杂的系统工程拆解为清晰的步骤，配合详细的图表与示例，让构建一个虽小但功能完备的大模型变得触手可及。无论你是想夯实理论基础，还是为未来研发更大规模的模型做准备",90106,"2026-04-06T11:19:32",[35,15,13,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":73,"owner_avatar_url":74,"owner_bio":75,"owner_company":76,"owner_location":76,"owner_email":77,"owner_twitter":78,"owner_website":79,"owner_url":80,"languages":81,"stars":119,"forks":120,"last_commit_at":121,"license":122,"difficulty_score":32,"env_os":123,"env_gpu":124,"env_ram":125,"env_deps":126,"category_tags":134,"github_topics":136,"view_count":32,"oss_zip_url":76,"oss_zip_packed_at":76,"status":17,"created_at":151,"updated_at":152,"faqs":153,"releases":184},5148,"uxlfoundation\u002FoneDAL","oneDAL","oneAPI Data Analytics Library (oneDAL)","oneDAL（oneAPI Data Analytics Library）是一个专为表格数据设计的高性能机器学习加速库，支持 C++ 和 DPC++ 语言。它内置了线性回归、K-means 聚类、随机森林等常用算法，能够显著提升数据处理速度。\n\n在数据分析中，面对海量表格数据时，传统算法往往运行缓慢，成为效率瓶颈。oneDAL 通过深度优化底层计算，有效解决了这一痛点。它在 CPU 上利用 SIMD 指令集和现代缓存结构进行加速，在 GPU 上则依托 SYCL 框架和 oneMKL 库释放硬件潜能，同时还能轻松扩展到多节点分布式环境，实现大规模并行计算。\n\n这款工具非常适合需要处理大规模数据的开发者、数据科学家以及研究人员。如果你正在使用 Python 的 scikit-learn 库，可以通过其专用扩展插件无缝调用 oneDAL，无需大幅修改代码即可让现有程序“提速”。对于追求极致性能的 C++ 开发者，oneDAL 也提供了灵活的底层接口。作为 UXL 基金会的一部分，oneDAL 致力于推动跨架构的统一开发标准，帮助你在不同硬件平台上构建高效、可移植的数据分析应用。","\u003C!--\n******************************************************************************\n* Copyright 2014 Intel Corporation\n*\n* Licensed under the Apache License, Version 2.0 (the \"License\");\n* you may not use this file except in compliance with the License.\n* You may obtain a copy of the License at\n*\n*     http:\u002F\u002Fwww.apache.org\u002Flicenses\u002FLICENSE-2.0\n*\n* Unless required by applicable law or agreed to in writing, software\n* distributed under the License is distributed on an \"AS IS\" BASIS,\n* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n* See the License for the specific language governing permissions and\n* limitations under the License.\n*******************************************************************************\u002F-->\n\n# oneAPI Data Analytics Library \u003C!-- omit in toc --> \u003Cimg align=\"right\" width=\"200\" height=\"100\" src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fuxlfoundation\u002Fartwork\u002Fe98f1a7a3d305c582d02c5f532e41487b710d470\u002Ffoundation\u002Fuxl-foundation-logo-horizontal-color.svg\">\n\n[Installation](#installation)   |   [Documentation](#documentation)   |   [Support](#support)   |   [Examples](#examples)   |   [How to Contribute](CONTRIBUTING.md)   \n\n[![Build Status](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_3323fe13e747.png)](https:\u002F\u002Fdev.azure.com\u002Fdaal\u002FDAAL\u002F_build\u002Flatest?definitionId=7&branchName=main)\n[![License](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fuxlfoundation\u002FoneDAL.svg)](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fblob\u002Fmain\u002FLICENSE)\n[![OpenSSF Best Practices](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_50ccde67b228.png)](https:\u002F\u002Fwww.bestpractices.dev\u002Fprojects\u002F8859)\n[![OpenSSF Scorecard](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_3fe773fe7c79.png)](https:\u002F\u002Fsecurityscorecards.dev\u002Fviewer\u002F?uri=github.com\u002Fuxlfoundation\u002FoneDAL)\n[![Join the community on GitHub Discussions](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_ca579e314549.png)](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fdiscussions)\n\noneAPI Data Analytics Library (oneDAL) is a C++ and DPC++ library (powering the [Extension for Scikit-learn in Python](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002Fscikit-learn-intelex))\nwhich implements accelerated machine learning routines for tabular data (e.g. linear regression, K-means clustering, random forests, etc.) for CPUs, GPUs, and\nmulti-node distributed setups.\n\nAcceleration on CPUs is achieved by leveraging SIMD instructions and exploiting cache structures of modern hardware, while GPU acceleration leverages the SYCL framework and the oneMKL library.\n\nOneDAL is part of the [UXL Foundation](http:\u002F\u002Fwww.uxlfoundation.org) and is an implementation of the [oneAPI specification](https:\u002F\u002Foneapi-spec.uxlfoundation.org) for the oneDAL component.\n\n## Usage\n\nThere are different ways for you to build high-performance data science applications that use the advantages of oneDAL:\n- Use [Extension for Scikit-learn*](https:\u002F\u002Fuxlfoundation.github.io\u002Fscikit-learn-intelex\u002F) to accelerate existing scikit-learn code by making it call oneDAL behind the scenes.\n- Use oneDAL C++ interfaces with or without SYCL support ([learn more](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002F#oneapi-vs-daal-interfaces)).\n\n\n## Installation\n\nCheck the [System Requirements](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002Fsystem-requirements.html) before installing to ensure compatibility with your system.\n\nThere are several options available for installing oneDAL:\n\n- **Binary Distribution**: pre-built binary packages are available from the following sources:\n    - Intel® oneAPI:\n        - Download as the Stand-Alone [oneAPI Data Analytics Library](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fdeveloper\u002Ftools\u002Foneapi\u002Fonedal-download.html)\n    - Conda:\n        | Channel | Version |\n        |:-------:|:-------:|\n        | conda-forge | [![Anaconda-Server Conda-forge Badge](https:\u002F\u002Fanaconda.org\u002Fconda-forge\u002Fdal-devel\u002Fbadges\u002Fversion.svg)](https:\u002F\u002Fanaconda.org\u002Fconda-forge\u002Fdal-devel) |\n\n    - [NuGet](https:\u002F\u002Fwww.nuget.org\u002Fpackages\u002Finteldal.devel.linux-x64)\n\n- **Source Distribution**: Clone this GitHub repository or [download a specific version of oneDAL](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Freleases) from the GitHub releases page and follow the instructions in the [INSTALL.md](INSTALL.md) file.\n\n\n## Examples\n\nC++ Examples:\n\n- [oneAPI interfaces with SYCL support](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fexamples\u002Foneapi\u002Fdpc)\n- [oneAPI interfaces without SYCL support](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fexamples\u002Foneapi\u002Fcpp)\n- [DAAL interfaces](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fexamples\u002Fdaal\u002Fcpp)\n\nPython Examples:\n- [scikit-learn-intelex](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002Fscikit-learn-intelex\u002Ftree\u002Fmain\u002Fexamples\u002Fnotebooks)\n\n\u003Cdetails>\u003Csummary>Other Examples\u003C\u002Fsummary>\n\n- [MPI](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fsamples\u002Fdaal\u002Fcpp\u002Fmpi)\n\n\u003C\u002Fdetails>\n\n## Documentation\n\noneDAL documentation:\n\n- [Release Notes](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Freleases)\n- [Get Started Guide](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002Fquick-start.html)\n- [Developer Guide and Reference](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002F)\n\nOther related documentation:\n\n- [Extension for Scikit-learn* documentation](https:\u002F\u002Fuxlfoundation.github.io\u002Fscikit-learn-intelex\u002F)\n- [oneDAL Specifications](https:\u002F\u002Foneapi-spec.uxlfoundation.org\u002Fspecifications\u002Foneapi\u002Flatest\u002Felements\u002Fonedal\u002Fsource\u002F#onedal-section)\n\n## Apache Spark MLlib\n\noneDAL library is used for Spark MLlib acceleration as part of [OAP MLlib](https:\u002F\u002Fgithub.com\u002Foap-project\u002Foap-mllib) project and allows you to get a **3-18x** increase in performance compared to the default Apache Spark MLlib.\n\n\u003Cimg style=\"display:inline;\" height=300 width=550 src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_cfbb0638b907.png\">\u003C\u002Fa>\n\n>*Technical details: FPType: double; HW: 7 x m5.2xlarge AWS instances; SW: Intel DAAL 2020 Gold, Apache Spark 2.4.4, emr-5.27.0; Spark config num executors 12, executor cores 8, executor memory 19GB, task cpus 8*\n\n## Scaling\n\noneDAL supports distributed computation mode that shows excellent results for strong and weak scaling:\n\noneDAL K-Means fit, strong scaling result | oneDAL K-Means fit, weak scaling results\n:-------------------------:|:-------------------------:\n![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_eab90f5d1177.png)  |   ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_70af54f1db4c.png)\n\n>*Technical details: FPType: float32; HW: Intel Xeon Processor E5-2698 v3 @2.3GHz, 2 sockets, 16 cores per socket; SW: Intel® DAAL (2019.3), MPI4Py (3.0.0), Intel® Distribution Of Python (IDP) 3.6.8; Details available in the article https:\u002F\u002Farxiv.org\u002Fabs\u002F1909.11822*\n\n## Governance\n\nThe oneDAL project is governed by the UXL Foundation and you can get involved in this project in multiple ways. It is possible to join the [AI Special Interest Group (SIG)](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002Ffoundation\u002Ftree\u002Fmain\u002Fai) meetings where the group discuss and demonstrates work using this project. Members can also join the Open Source and Specification Working Group meetings.\n\nYou can also join the mailing lists for the [UXL Foundation](https:\u002F\u002Flists.uxlfoundation.org\u002Fg\u002Fmain\u002Fsubgroups) to be informed of when meetings are happening and receive the latest information and discussions.\n\nYou can contribute to this project and also contribute to the specification for this project, read the [CONTRIBUTING](CONTRIBUTING.md) page for more information.\n\n\n## Support\n\nAsk questions and engage in discussions with oneDAL developers, contributers, and other users through the following channels:\n\n- [GitHub Discussions](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fdiscussions)\n- [Community Forum](https:\u002F\u002Fcommunity.intel.com\u002Ft5\u002FIntel-oneAPI-Data-Analytics\u002Fbd-p\u002Foneapi-data-analytics-library)\n\nYou may reach out to project maintainers privately at onedal.maintainers@intel.com.\n\n### Security \u003C!-- omit in toc -->\n\nTo report a vulnerability, refer to [Intel vulnerability reporting policy](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fsecurity-center\u002Fdefault.html).\n\n### Contribute \u003C!-- omit in toc -->\n\nWe welcome community contributions. Check our [contributing guidelines](CONTRIBUTING.md) to learn more. You can also contact the oneDAL team via [UXL Foundation Slack] using\n[#onedal] channel.\n\n[UXL Foundation Slack]: https:\u002F\u002Fslack-invite.uxlfoundation.org\u002F\n[#onedal]: https:\u002F\u002Fuxlfoundation.slack.com\u002Fchannels\u002Fonedal\n\n## License \u003C!-- omit in toc -->\n\noneDAL is distributed under the Apache License 2.0 license. See [LICENSE](LICENSE) for more information.\n","\u003C!--\n******************************************************************************\n* 版权所有 © 2014 英特尔公司\n*\n* 根据 Apache License, Version 2.0（“许可证”）获得许可；\n* 除非符合许可证的规定，否则不得使用本文件。\n* 您可以在以下地址获取许可证副本：\n*\n*     http:\u002F\u002Fwww.apache.org\u002Flicenses\u002FLICENSE-2.0\n*\n* 除非适用法律要求或双方另行约定，否则软件按“原样”分发，\n* 不提供任何形式的保证或条件，无论是明示的还是默示的。\n* 许可证详细规定了权限和限制。\n*******************************************************************************\u002F-->\n\n# oneAPI 数据分析库 \u003C!-- omit in toc --> \u003Cimg align=\"right\" width=\"200\" height=\"100\" src=\"https:\u002F\u002Fraw.githubusercontent.com\u002Fuxlfoundation\u002Fartwork\u002Fe98f1a7a3d305c582d02c5f532e41487b710d470\u002Ffoundation\u002Fuxl-foundation-logo-horizontal-color.svg\">\n\n[安装](#installation)   |   [文档](#documentation)   |   [支持](#support)   |   [示例](#examples)   |   [如何贡献](CONTRIBUTING.md)   \n\n[![构建状态](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_3323fe13e747.png)](https:\u002F\u002Fdev.azure.com\u002Fdaal\u002FDAAL\u002F_build\u002Flatest?definitionId=7&branchName=main)\n[![许可证](https:\u002F\u002Fimg.shields.io\u002Fgithub\u002Flicense\u002Fuxlfoundation\u002FoneDAL.svg)](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fblob\u002Fmain\u002FLICENSE)\n[![OpenSSF 最佳实践](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_50ccde67b228.png)](https:\u002F\u002Fwww.bestpractices.dev\u002Fprojects\u002F8859)\n[![OpenSSF 安全评分卡](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_3fe773fe7c79.png)](https:\u002F\u002Fsecurityscorecards.dev\u002Fviewer\u002F?uri=github.com\u002Fuxlfoundation\u002FoneDAL)\n[![加入 GitHub Discussions 社区](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_ca579e314549.png)](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fdiscussions)\n\noneAPI 数据分析库（oneDAL）是一个 C++ 和 DPC++ 库（为 Python 中的 [Scikit-learn 扩展](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002Fscikit-learn-intelex) 提供支持），它实现了针对表格数据的加速机器学习算法（例如线性回归、K 均值聚类、随机森林等），适用于 CPU、GPU 以及多节点分布式环境。\n\n在 CPU 上的加速是通过利用 SIMD 指令和现代硬件的缓存结构来实现的，而 GPU 加速则依赖于 SYCL 框架和 oneMKL 库。\n\nOneDAL 是 [UXL 基金会](http:\u002F\u002Fwww.uxlfoundation.org) 的一部分，也是 [oneAPI 规范](https:\u002F\u002Foneapi-spec.uxlfoundation.org) 在 oneDAL 组件中的具体实现。\n\n## 使用方法\n\n您可以通过多种方式构建利用 oneDAL 优势的高性能数据科学应用程序：\n- 使用 [Scikit-learn* 扩展](https:\u002F\u002Fuxlfoundation.github.io\u002Fscikit-learn-intelex\u002F) 来加速现有的 scikit-learn 代码，使其在后台调用 oneDAL。\n- 使用 oneDAL 的 C++ 接口，无论是否支持 SYCL（[了解更多](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002F#oneapi-vs-daal-interfaces)）。\n\n\n## 安装\n\n在安装之前，请查看 [系统要求](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002Fsystem-requirements.html)，以确保与您的系统兼容。\n\n有几种可用的 oneDAL 安装选项：\n\n- **二进制分发**：预编译的二进制包可从以下来源获取：\n    - Intel® oneAPI：\n        - 可作为独立的 [oneAPI 数据分析库](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fdeveloper\u002Ftools\u002Foneapi\u002Fonedal-download.html) 下载\n    - Conda：\n        | 渠道 | 版本 |\n        |:-------:|:-------:|\n        | conda-forge | [![Anaconda-Server Conda-forge 徽章](https:\u002F\u002Fanaconda.org\u002Fconda-forge\u002Fdal-devel\u002Fbadges\u002Fversion.svg)](https:\u002F\u002Fanaconda.org\u002Fconda-forge\u002Fdal-devel) |\n\n    - [NuGet](https:\u002F\u002Fwww.nuget.org\u002Fpackages\u002Finteldal.devel.linux-x64)\n\n- **源码分发**：克隆此 GitHub 仓库，或从 GitHub 发布页面 [下载特定版本的 oneDAL](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Freleases)，并按照 [INSTALL.md](INSTALL.md) 文件中的说明进行操作。\n\n\n## 示例\n\nC++ 示例：\n\n- [支持 SYCL 的 oneAPI 接口](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fexamples\u002Foneapi\u002Fdpc)\n- [不支持 SYCL 的 oneAPI 接口](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fexamples\u002Foneapi\u002Fcpp)\n- [DAAL 接口](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fexamples\u002Fdaal\u002Fcpp)\n\nPython 示例：\n- [scikit-learn-intelex](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002Fscikit-learn-intelex\u002Ftree\u002Fmain\u002Fexamples\u002Fnotebooks)\n\n\u003Cdetails>\u003Csummary>其他示例\u003C\u002Fsummary>\n\n- [MPI](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fsamples\u002Fdaal\u002Fcpp\u002Fmpi)\n\n\u003C\u002Fdetails>\n\n## 文档\n\noneDAL 文档：\n\n- [发布说明](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Freleases)\n- [入门指南](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002Fquick-start.html)\n- [开发者指南和参考手册](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002F)\n\n其他相关文档：\n\n- [Scikit-learn* 扩展文档](https:\u002F\u002Fuxlfoundation.github.io\u002Fscikit-learn-intelex\u002F)\n- [oneDAL 规范](https:\u002F\u002Foneapi-spec.uxlfoundation.org\u002Fspecifications\u002Foneapi\u002Flatest\u002Felements\u002Fonedal\u002Fsource\u002F#onedal-section)\n\n## Apache Spark MLlib\n\noneDAL 库被用于 Spark MLlib 的加速，作为 [OAP MLlib](https:\u002F\u002Fgithub.com\u002Foap-project\u002Foap-mllib) 项目的一部分，与默认的 Apache Spark MLlib 相比，性能可提升 **3 至 18 倍**。\n\n\u003Cimg style=\"display:inline;\" height=300 width=550 src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_cfbb0638b907.png\">\u003C\u002Fa>\n\n>*技术细节：FPType：double；硬件：7 台 m5.2xlarge AWS 实例；软件：Intel DAAL 2020 Gold、Apache Spark 2.4.4、emr-5.27.0；Spark 配置：执行器数量 12、每个执行器核心数 8、每个执行器内存 19GB、任务 CPU 数 8*\n\n## 扩展性\n\noneDAL 支持分布式计算模式，在强扩展性和弱扩展性方面均表现出色：\n\noneDAL K-Means 拟合，强扩展性结果 | oneDAL K-Means 拟合，弱扩展性结果\n:-------------------------:|:-------------------------:\n![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_eab90f5d1177.png)  |   ![](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_readme_70af54f1db4c.png)\n\n>*技术细节：FPType：float32；硬件：Intel Xeon 处理器 E5-2698 v3 @2.3GHz，双插槽，每插槽 16 核；软件：Intel® DAAL (2019.3)、MPI4Py (3.0.0)、Intel® Distribution Of Python (IDP) 3.6.8；详情请参阅论文 https:\u002F\u002Farxiv.org\u002Fabs\u002F1909.11822*\n\n## 治理\n\noneDAL 项目由 UXL 基金会管理，您可以通过多种方式参与其中。您可以加入 [AI 特别兴趣小组 (SIG)](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002Ffoundation\u002Ftree\u002Fmain\u002Fai) 的会议，在这些会议上，小组成员会讨论并演示使用该项目的工作成果。成员也可以参加开源与规范工作组的会议。\n\n此外，您还可以加入 [UXL 基金会](https:\u002F\u002Flists.uxlfoundation.org\u002Fg\u002Fmain\u002Fsubgroups) 的邮件列表，以便及时了解会议安排，并获取最新资讯和讨论内容。\n\n您既可以为本项目做出贡献，也可以参与该项目的规范制定。更多信息请参阅 [CONTRIBUTING](CONTRIBUTING.md) 页面。\n\n\n## 支持\n\n您可以通过以下渠道向 oneDAL 开发者、贡献者及其他用户提问并参与讨论：\n\n- [GitHub Discussions](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fdiscussions)\n- [社区论坛](https:\u002F\u002Fcommunity.intel.com\u002Ft5\u002FIntel-oneAPI-Data-Analytics\u002Fbd-p\u002Foneapi-data-analytics-library)\n\n如需私下联系项目维护人员，请发送邮件至 onedal.maintainers@intel.com。\n\n### 安全 \u003C!-- omit in toc -->\n\n如需报告漏洞，请参考 [英特尔漏洞报告政策](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fsecurity-center\u002Fdefault.html)。\n\n### 贡献 \u003C!-- omit in toc -->\n\n我们欢迎社区贡献。请查阅我们的 [贡献指南](CONTRIBUTING.md)，了解更多详情。您也可以通过 [UXL 基金会 Slack](https:\u002F\u002Fslack-invite.uxlfoundation.org\u002F) 的 [#onedal] 频道联系 oneDAL 团队。\n\n[UXL 基金会 Slack]: https:\u002F\u002Fslack-invite.uxlfoundation.org\u002F\n[#onedal]: https:\u002F\u002Fuxlfoundation.slack.com\u002Fchannels\u002Fonedal\n\n## 许可证 \u003C!-- omit in toc -->\n\noneDAL 根据 Apache License 2.0 许可证进行分发。更多信息请参阅 [LICENSE](LICENSE) 文件。","# oneDAL 快速上手指南\n\noneAPI Data Analytics Library (oneDAL) 是一个高性能的 C++ 和 DPC++ 库，专为表格数据加速机器学习算法（如线性回归、K-means 聚类、随机森林等）。它支持 CPU（利用 SIMD 指令）、GPU（基于 SYCL）以及多节点分布式环境。\n\n## 1. 环境准备\n\n在开始之前，请确保您的系统满足以下要求：\n\n*   **操作系统**：Linux (Ubuntu, CentOS, RHEL 等), Windows, macOS。\n*   **编译器**：\n    *   **推荐**：Intel® oneAPI DPC++\u002FC++ Compiler (用于完整功能及 GPU 加速)。\n    *   **兼容**：GCC, Clang, MSVC (仅用于 CPU 基础功能)。\n*   **依赖项**：\n    *   若需 GPU 加速，需安装支持 SYCL 的后端（如 Intel GPU Driver 和 oneAPI Level Zero）。\n    *   若需分布式计算，需安装 MPI (如 Intel MPI, OpenMPI)。\n*   **Python 用户**：若希望通过 Python 使用，建议安装 `scikit-learn`，oneDAL 可通过 `scikit-learn-intelex` 对其进行透明加速。\n\n> **注意**：详细系统要求请参考 [官方文档](https:\u002F\u002Fuxlfoundation.github.io\u002FoneDAL\u002Fsystem-requirements.html)。\n\n## 2. 安装步骤\n\n您可以根据开发需求选择以下任一方式安装：\n\n### 方式 A：通过 Conda 安装（推荐 Python\u002FC++ 开发者）\n\n这是最便捷的方式，适用于大多数 Linux 和 Windows 环境。\n\n```bash\nconda install -c conda-forge dal-devel\n```\n\n### 方式 B：通过 Intel oneAPI 工具包安装（适合需要完整生态的用户）\n\n访问 [Intel oneAPI 下载页面](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fdeveloper\u002Ftools\u002Foneapi\u002Fonedal-download.html) 下载 standalone 安装包，或使用命令行工具安装：\n\n```bash\n# 示例：使用 apt (Ubuntu\u002FDebian) 添加源并安装\nwget -O- https:\u002F\u002Fapt.repos.intel.com\u002Fintel-gpg-keys\u002FGPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB | gpg --dearmor | sudo tee \u002Fusr\u002Fshare\u002Fkeyrings\u002Foneapi-archive-keyring.gpg > \u002Fdev\u002Fnull\necho \"deb [signed-by=\u002Fusr\u002Fshare\u002Fkeyrings\u002Foneapi-archive-keyring.gpg] https:\u002F\u002Fapt.repos.intel.com\u002Foneapi all main\" | sudo tee \u002Fetc\u002Fapt\u002Fsources.list.d\u002FoneAPI.list\nsudo apt update\nsudo apt install intel-oneapi-daal\n```\n\n*注：国内用户若遇到下载速度慢的问题，可尝试配置清华或中科大镜像源代理 Conda 通道，或手动下载离线包安装。*\n\n### 方式 C：源码编译（适合高级开发者）\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL.git\ncd oneDAL\n# 请参考根目录下的 INSTALL.md 文件进行具体编译配置\n```\n\n## 3. 基本使用\n\noneDAL 主要有两种使用模式：**Python 透明加速**（最简单）和 **C++ 原生调用**。\n\n### 场景一：Python 用户（加速现有 Scikit-learn 代码）\n\n无需修改算法逻辑，只需导入 `daal4py` 或启用 `scikit-learn-intelex` 即可自动将后端切换为 oneDAL。\n\n**安装扩展包：**\n```bash\npip install scikit-learn-intelex\n```\n\n**使用示例：**\n```python\nfrom sklearnex import patch_sklearn\npatch_sklearn()\n\n# 现在所有的 scikit-learn 调用都将自动使用 oneDAL 加速\nfrom sklearn.datasets import make_classification\nfrom sklearn.ensemble import RandomForestClassifier\n\nX, y = make_classification(n_samples=1000, n_features=20)\nclf = RandomForestClassifier(n_estimators=100)\nclf.fit(X, y)\n\nprint(\"训练完成，已自动加速\")\n```\n\n### 场景二：C++ 用户（原生接口）\n\n以下是一个使用 oneDAL C++ 接口进行 K-Means 聚类的最小化示例（需链接 oneDAL 库）。\n\n```cpp\n#include \"oneapi\u002Fdal.hpp\"\n#include \u003Ciostream>\n\nint main() {\n    \u002F\u002F 1. 准备数据 (此处仅为示意，实际需填充数据)\n    auto data = oneapi::dal::homogen_table::empty(); \n    \u002F\u002F ... 加载数据到 data ...\n\n    \u002F\u002F 2. 描述算法参数\n    auto desc = oneapi::dal::kmeans::descriptor\u003Cfloat>{\n        3, \u002F\u002F 聚类中心数量\n        oneapi::dal::kmeans::method::lloyd,\n        oneapi::dal::kmeans::init::random_dense\n    };\n\n    try {\n        \u002F\u002F 3. 执行训练\n        auto result = oneapi::dal::train(desc, data);\n\n        \u002F\u002F 4. 获取结果\n        auto centers = result.get_model().get_centroids();\n        std::cout \u003C\u003C \"K-Means 训练完成，中心点行数：\" \u003C\u003C centers.get_row_count() \u003C\u003C std::endl;\n    } catch (const std::exception& e) {\n        std::cerr \u003C\u003C \"Error: \" \u003C\u003C e.what() \u003C\u003C std::endl;\n        return 1;\n    }\n\n    return 0;\n```\n\n**编译命令示例 (使用 icpx):**\n```bash\nicpx -std=c++17 -I${ONEAPI_ROOT}\u002Fdal\u002Flatest\u002Finclude your_code.cpp -L${ONEAPI_ROOT}\u002Fdal\u002Flatest\u002Flib\u002Fintel64 -ldal -o kmeans_app\n```\n\n---\n更多详细示例代码请访问 [GitHub Examples](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Fmain\u002Fexamples)。","某大型电商公司的数据科学团队正致力于构建实时用户行为分析系统，需要对海量日志数据进行快速的聚类分析和异常检测，以支持动态推荐策略。\n\n### 没有 oneDAL 时\n- **训练耗时过长**：面对亿级行数的用户点击流数据，使用原生 Scikit-learn 进行 K-means 聚类往往需要数小时甚至更久，无法满足“准实时”的业务需求。\n- **硬件资源闲置**：算法仅能利用 CPU 的单核或少量核心，无法有效调用服务器配备的高性能 GPU 或多节点集群算力，导致昂贵的硬件资源大量浪费。\n- **代码重构成本高**：若要引入其他加速库，通常需要大幅重写现有的 Python 数据分析代码，不仅开发周期长，还容易引入新的 Bug。\n- **扩展性受限**：随着数据量激增，单机内存和算力迅速达到瓶颈，难以平滑扩展到分布式环境处理更大规模的数据集。\n\n### 使用 oneDAL 后\n- **速度显著提升**：通过 scikit-learn-intelex 无缝接入 oneDAL，利用其优化的 SIMD 指令和缓存机制，将原本数小时的聚类任务缩短至几分钟内完成。\n- **异构算力全开**：自动调度 CPU、GPU 及多节点资源，充分发挥现代硬件的并行计算能力，使吞吐量提升数十倍。\n- **零代码迁移**：无需修改任何业务逻辑代码，仅需在导入环节增加一行配置，即可让现有的 Scikit-learn 模型后台调用 oneDAL 加速引擎。\n- **弹性伸缩自如**：轻松应对从单机到多节点分布式部署的平滑过渡，支撑起 PB 级数据的实时分析需求，保障业务高峰期的稳定性。\n\noneDAL 的核心价值在于让开发者无需重写代码，即可将传统机器学习算法的性能释放到极致，真正实现“一次编写，处处加速”。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fuxlfoundation_oneDAL_cfbb0638.png","uxlfoundation","Unified Acceleration (UXL) Foundation","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fuxlfoundation_bc0c99d1.png","Building a multi-architecture and multi-vendor software ecosystem for all accelerators",null,"operations@uxlfoundation.org","UXLfoundation","https:\u002F\u002Fuxlfoundation.org","https:\u002F\u002Fgithub.com\u002Fuxlfoundation",[82,86,89,93,97,101,105,109,112,116],{"name":83,"color":84,"percentage":85},"C++","#f34b7d",83.4,{"name":87,"color":76,"percentage":88},"SWIG",13.1,{"name":90,"color":91,"percentage":92},"Starlark","#76d275",1.8,{"name":94,"color":95,"percentage":96},"Makefile","#427819",0.9,{"name":98,"color":99,"percentage":100},"Shell","#89e051",0.4,{"name":102,"color":103,"percentage":104},"C","#555555",0.3,{"name":106,"color":107,"percentage":108},"CMake","#DA3434",0.1,{"name":110,"color":111,"percentage":108},"Batchfile","#C1F12E",{"name":113,"color":114,"percentage":115},"Dockerfile","#384d54",0,{"name":117,"color":118,"percentage":115},"Smarty","#f0c040",645,224,"2026-04-04T13:27:57","Apache-2.0","Linux, Windows","非必需。支持通过 SYCL 框架和 oneMKL 库进行 GPU 加速（未明确指定具体显卡型号或显存要求，通常指支持 SYCL 的 Intel GPU 或其他兼容设备）。","未说明",{"notes":127,"python":128,"dependencies":129},"该库主要提供 C++ 和 DPC++ 接口，也可通过 Scikit-learn 扩展在 Python 中使用。CPU 加速利用 SIMD 指令和现代硬件缓存结构。支持多节点分布式设置。详细系统要求需查阅官方文档链接。NuGet 包明确支持 linux-x64。","未说明（提供 Scikit-learn 扩展包，具体版本需参考该扩展包要求）",[130,131,132,133],"oneMKL (用于 GPU 加速)","SYCL (用于 GPU 编程)","MPI (用于多节点分布式计算)","Scikit-learn (可选，用于 Python 接口)",[16,52,135,14],"其他",[137,138,139,140,141,142,143,144,145,146,147,148,149,150],"onedal","data-analysis","oneapi","machine-learning","machine-learning-algorithms","hacktoberfest","swrepo","ai-machine-learning","big-data","analytics","ai-training","ai-inference","cpp","data-science","2026-03-27T02:49:30.150509","2026-04-08T01:08:46.559458",[154,159,164,169,174,179],{"id":155,"question_zh":156,"answer_zh":157,"source_url":158},23357,"如何在 FreeBSD 上编译和移植 oneDAL？","oneDAL 自 2019 Update 1 版本起已正式支持 FreeBSD。如果遇到编译问题，通常只需进行少量修改即可编译库和示例。您可以参考相关的 Pull Request（如 PR #53）获取具体的代码变更建议。确保使用支持 FreeBSD 的版本后，按照常规步骤编译即可正常工作。","https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fissues\u002F45",{"id":160,"question_zh":161,"answer_zh":162,"source_url":163},23358,"是否支持在 C++ 中使用由 Python (LightGBM\u002FXGBoost) 训练的模型进行推理？","目前仓库中没有直接对应“Python 训练、C++ 推理”这一特定流程的示例格式。但是，您可以参考 daal4py 的相关示例来了解模型转换逻辑：https:\u002F\u002Fgithub.com\u002Fintel\u002Fscikit-learn-intelex\u002Ftree\u002Fmaster\u002Fexamples\u002Fdaal4py。维护团队正在考虑为此特定场景准备演示代码。当前建议是先研究现有的 Python 示例以理解模型导出格式。","https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fissues\u002F1530",{"id":165,"question_zh":166,"answer_zh":167,"source_url":168},23359,"如何在 HiBench 中构建并运行基于 DAAL 的 Spark KMeans 测试？","DAAL KMeans 已正式集成到 HiBench 中。构建时需添加 `-Dmodules -Pdal` 参数，例如：\n`mvn -Psparkbench -Dmodules -Pml -Pdal -Dspark=2.0 -Dscala=2.11 clean package`\n\n运行前请安装 DAL 并设置所需的环境变量。然后使用以下脚本生成数据并运行任务：\n1. 准备数据：`bin\u002Fworkloads\u002Fdal\u002Fkmeans\u002Fprepare\u002Fprepare.sh`\n2. 运行任务：`bin\u002Fworkloads\u002Fdal\u002Fkmeans\u002Fspark\u002Frun.sh`\n\n注意确保分区大小配置合理，以避免潜在的崩溃问题。","https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fissues\u002F174",{"id":170,"question_zh":171,"answer_zh":172,"source_url":173},23360,"makefile 不支持交叉编译（如 Yocto 框架）怎么办？","早期的 makefile 硬编码了 `g++` 和 `ar`，导致交叉编译失败（报错 \"File format not recognized\"）。该问题已通过 PR 修复。修复后的 makefile 会使用隐式变量 `${CXX}` 和 `${AR}` 替代硬编码编译器，从而兼容 GNU 标准的交叉编译环境。如果您仍遇到此问题，请确保拉取了包含该修复的最新代码。","https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fissues\u002F57",{"id":175,"question_zh":176,"answer_zh":177,"source_url":178},23361,"不同分类器（如 Decision Forest 和 KNN）的构造函数接口不一致，如何处理类数量参数？","针对接口不一致的问题（部分算法需要 `nClasses` 参数，部分不需要），维护者已在主分支中更新了 kNN 的构造函数。现在新的 kNN 构造函数也支持传入 `nClasses` 参数，以提高与其他分类器（如 Decision Forest）的一致性。请升级到包含 PR #1392 或更新版本的代码库来解决此编译错误。","https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fissues\u002F299",{"id":180,"question_zh":181,"answer_zh":182,"source_url":183},23362,"如何正确序列化和反序列化逻辑回归、决策森林等分类模型？","直接使用 `new Model()` 创建模型指针会导致编译错误，因为这些具体模型的 `Model` 类是抽象类。正确的做法是使用训练结果对象（ResultPtr）来获取模型指针，而不是尝试实例化抽象的 Model 类。即：先获取 `training::result`，然后从中提取 `model` 进行序列化操作。这是处理逻辑回归、梯度提升树和决策森林等模型序列化的预期方式。","https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fissues\u002F263",[185,190,195,200,205,210,215,220,225,230,235,240,245,250,255,260,265,270,275,280],{"id":186,"version":187,"summary_zh":188,"released_at":189},144857,"2025.11.0","oneDAL很高兴地推出2025.11.0版本！\n\n## :beetle: 错误修复\n\n* 修复了CPU拓扑结构初始化错误\n\n## :octocat: 致谢\n\n感谢所有帮助我们实现2025.11.0版本发布的朋友们！\n\n@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel、@yuejiaointel、@DDJHB、@kjackiew、@richardnorth3\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.10.1...2025.11.0","2026-03-09T11:16:19",{"id":191,"version":192,"summary_zh":193,"released_at":194},144858,"2025.10.1","oneDAL 很高兴地宣布 2025.10.1 版本发布！\n\n## 变更内容\n* 该版本是 [Scikit-learn 扩展库* 2025.10.1](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002Fscikit-learn-intelex\u002Freleases\u002Ftag\u002F2025.10.1) 的支持版本。\n\n感谢所有帮助我们实现 2025.10.1 版本发布的人员！\n\n@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel、@yuejiaointel、@DDJHB、@kjackiew、@richardnorth3","2026-01-23T15:27:18",{"id":196,"version":197,"summary_zh":198,"released_at":199},144865,"2025.4.0","oneDAL 很高兴地推出 2025.4.0 版本！\n\n## :rotating_light: 新特性\n\n* 引入了 oneDAL 的新功能：\n  * 对 oneDAL 随机数生成器原语函数的统一 API 进行了更新\n  * 更改了随机数生成器克隆操作中的流处理方式\n\n## :beetle: 错误修复\n\n* 修复了 DAAL 代码示例的相关问题（使用 C++20 编译时）\n\n## :octocat: 致谢\n\n感谢所有帮助我们实现 2025.4.0 版本发布的朋友们！\n\n@Alexsandruss、@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.2.0...2025.4.0","2025-04-02T12:48:46",{"id":201,"version":202,"summary_zh":203,"released_at":204},144866,"2025.2.0","# oneDAL 2025.2.0\noneDAL 非常高兴地推出 2025.2.0 版本！\n\n## :rotating_light: 新特性\n\n* 引入了新的 oneDAL 功能：\n  * 实现了作为线性回归后备方案的 GPU 版本谱分解\n  * 增加了 VTune 性能分析功能\n\n## :beetle: 错误修复\n\n* 修复了当桶数少于作业数时的错误归约操作\n* 修复了 AlgorithmContainer 类中的“三法则”违规问题\n* 修复了主机到设备数据传输中的额外数组分配问题\n* 修复了低阶矩中的 `min`、`max` 和 `sum_squares` 计算问题\n* 修复了线程 ID 映射错误\n\n## :octocat: 致谢\n感谢所有帮助我们实现 2025.2.0 版本发布的朋友们！\n\n@Alexsandruss、@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.1.0...2025.2.0","2025-02-24T16:43:51",{"id":206,"version":207,"summary_zh":208,"released_at":209},144867,"2025.1.0","oneDAL is happy to introduce 2025.1.0 release!\r\n\r\n## :triangular_flag_on_post: Removals and ABI Compatibility \r\n\r\n* Removed `libonedal_sycl.a` from the distribution\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * Added OpenRNG Backend\r\n  * Added xcp reference implementation\r\n* Improved oneDAL performance for the following algorithms:\r\n  * SVE optimised float WSSJ kernel\r\n\r\n## :beetle: Bug Fixes\r\n\r\n* Addressed missing dependencies to compute probabilities function\r\n* Added missing headers for `finiteness_checker algorithm`\r\n* Fixed broken links in the documentation\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2025.1.0 release possible! \r\n\r\n@Alexsandruss, @Alexandr-Solovev, @Vika-F, @david-cortes-intel, @icfaust, @napetrov, @maria-Petrova, @homksei, @ahuber21, @ethanglaser, @samir-nasibli, @olegkkruglov, @razdoburdin, @avolkov-intel, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.0.0...2025.1.0","2025-01-17T12:19:52",{"id":211,"version":212,"summary_zh":213,"released_at":214},144875,"2024.1.0","# oneDAL 2024.1.0\r\noneDAL is happy to introduce 2024.1.0 release!\r\n\r\n## :rotating_light: What's New\r\n\r\n* New oneDAL functionality:\r\n  * Enabled distributed computations for LogisticRegression algorithm\r\n  * Basic statistics algorithm for sparse data\r\n* Added new parameters to oneDAL algorithms:\r\n  * Bias parameter to Covariance algorithm\r\n* Improved oneDAL performance for the following algorithms:\r\n  * DBSCAN\r\n  * Distributed version of kNN\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.1.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @aepanchi, @KulikovNikita, @icfaust, @inteldimitrius, @maria-Petrova, @samir-nasibli, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.0.1...2024.1.0\r\n","2024-01-24T21:43:54",{"id":216,"version":217,"summary_zh":218,"released_at":219},144876,"2024.0.1","# oneDAL 2024.0.1\r\noneDAL is happy to introduce 2024.0.1 release! \r\n\r\n## :rotating_light: What's New\r\n\r\n* New oneDAL functionality: \r\n  * Introduced Logistic Regression algorithm\r\n  * Introduced online algorithms support: Moments of Low Order, PCA, Linear Regression\r\n  * Introduced dispatching in oneAPI covariance algorithm\r\n* New Model Builders functionality: \r\n  * SHAP calculation is added to GBT regression\r\n\r\n## :books: Support Materials\r\n\r\n[Faster XGBoost*, LightGBM, and CatBoost Inference on the CPU](https:\u002F\u002Fwww.intel.com\u002Fcontent\u002Fwww\u002Fus\u002Fen\u002Fdeveloper\u002Farticles\u002Ftechnical\u002Ffaster-xgboost-light-gbm-catboost-inference-on-cpu.html#gs.1m2dxh)\r\n\r\n## :twisted_rightwards_arrows: Adoption\r\n\r\n[TPOT2 AutoML  integration](https:\u002F\u002Fgithub.com\u002FEpistasisLab\u002Ftpot2\u002Fpull\u002F102)\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.0.1 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @aepanchi, @KulikovNikita, @icfaust, @inteldimitrius, @maria-Petrova, @samir-nasibli, @md-shafiul-alam\r\n\r\n\r\n","2023-11-30T17:50:26",{"id":221,"version":222,"summary_zh":223,"released_at":224},144859,"2025.10.0","oneDAL 很高兴地推出 2025.10.0 版本！\n\n## :triangular_flag_on_post: 移除与 ABI 兼容性\n\n* 以下功能已弃用，将在 oneDAL 2026.0.0 版本中移除：\n  * [Matrix](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fblob\u002Fmain\u002Fcpp\u002Fdaal\u002Finclude\u002Fdata_management\u002Fdata\u002Fmatrix.h#L49)、[PackedSymmetricMatrix](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fblob\u002Fmain\u002Fcpp\u002Fdaal\u002Finclude\u002Fdata_management\u002Fdata\u002Fsymmetric_matrix.h#L106) 和 [PackedTriangularMatrix](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fblob\u002Fmain\u002Fcpp\u002Fdaal\u002Finclude\u002Fdata_management\u002Fdata\u002Fsymmetric_matrix.h#L801) 数据类型\n  * ODBC、KDB 和 MySQL 数据源\n  * 在 [20251015-DAAL-algorithms-deprecation](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Frfcs\u002Frfcs\u002F20251015-DAAL-algorithms-deprecation) RFC 中列出的算法将被完全弃用\n  * 在 [20251022-DAAL-interfaces-deprecation](https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Ftree\u002Frfcs\u002Frfcs\u002F20251022-DAAL-interfaces-deprecation) RFC 中列出的算法将从 DAAL API 中移除，但其功能仍将通过其他 API 提供\n\n## :rotating_light: 新特性\n\n* 引入了新的 oneDAL 功能：\n  * 向量化随机森林均值\u002F方差计算\n\n## :beetle: 错误修复\n\n* 修复了 VC 编译器的 Makefile 标志错误\n* 修复了当随机森林中存在零权重时出现的除以零问题\n* 修复了 DAAL 中的内存泄漏\n* 修复了 icx、GNU、clang 和 MS VC 编译器的向量化编译指示\n* 修复了 GCC 中的未定义行为警告\n\n## :octocat: 致谢\n\n感谢所有帮助我们实现 2025.10.0 版本发布的人员！\n\n@Alexsandruss、@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel、@yuejiaointel、@DDJHB、@kjackiew、@richardnorth3\n\n## 新贡献者\n* @avineet4 在 https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fpull\u002F3410 中做出了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.9.0...2025.10.0","2025-12-10T15:36:27",{"id":226,"version":227,"summary_zh":228,"released_at":229},144860,"2025.9.0","oneDAL 很高兴地推出 2025.9.0 版本！\n\n## :rotating_light: 新功能\n\n* 引入了 oneDAL 的新功能：\n  * 在 kNN 搜索中启用了 SPMD API 支持\n\n## :beetle: 错误修复\n\n* 修复了 `csr_table` 转换中的内存分配不足问题\n* 防止逻辑回归在预测时结果变为 `inf`\n\n## :octocat: 致谢\n\n感谢所有帮助我们实现 2025.9.0 版本发布的朋友们！\n\n@Alexsandruss、@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel、@KateBlueSky、@yuejiaointel、@DDJHB、@kjackiew、@richardnorth3\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.8.0...2025.9.0","2025-10-27T11:52:33",{"id":231,"version":232,"summary_zh":233,"released_at":234},144861,"2025.8.0","oneDAL 很高兴地推出 2025.8.0 版本！\n\n## :rotating_light: 新功能\n\n* 引入了 oneDAL 的新功能：\n  * 改进了逻辑函数的数值精度\n  * 添加了余弦距离算法\n  * 为 postGemmPart 函数添加了基于 SLEEF 的 SVE 优化浮点指数实现\n  * 提升了 kNN 预测的性能\n\n## :beetle: 错误修复\n\n* 修复了 SSE4.2 上的非法指令问题\n* 修复了 modelbuilders 示例中的错误\n* 逻辑损失现在避免了变为无穷大\n\n## :octocat: 致谢\n\n感谢所有帮助我们实现 2025.8.0 版本发布的朋友们！\n\n@Alexsandruss、@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel、@KateBlueSky、@yuejiaointel、@DDJHB、@kjackiew、@richardnorth3\n\n## 新贡献者\n\n* @shubhamsvc 在 https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fpull\u002F3271 中做出了他们的首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.7.0...2025.8.0","2025-08-20T08:08:03",{"id":236,"version":237,"summary_zh":238,"released_at":239},144862,"2025.7.0","oneDAL 很高兴地推出 2025.7.0 版本！\n\n## :rotating_light: 新特性\n\n* 引入了 oneDAL 的新功能：\n  * 改进了 Windows 平台上动态符号加载的线程安全性\n  * 为 daal 示例添加了 Bazel 支持\n  * 添加了相关距离算法\n  * 在协方差和 PCA 算法中新增了 `grain_size` 超参数\n\n## :beetle: 错误修复\n\n* 修复了 GNU 编译器下共享库与 AddressSanitizer 的集成问题\n* 修复了 `xcsrmv` 函数，以确保协方差算法的正确计算\n* 修复了协方差中的 `hyperparameterIdCount` 问题\n\n## :octocat: 致谢\n感谢所有帮助我们实现 2025.7.0 版本发布的朋友们！\n\n@Alexsandruss、@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.6.0...2025.7.0","2025-07-10T09:33:29",{"id":241,"version":242,"summary_zh":243,"released_at":244},144863,"2025.6.0","oneDAL 很高兴地宣布 2025.6.1 版本发布！\n\n## :triangular_flag_on_post: 移除与 ABI 兼容性\n\n* oneDAL 静态 SYCL 库在 2025.6.1 版本中已弃用并移除。\n\n## :hammer: 库工程改进\n\n* 已将 `CPATH` 变量替换为 `CPLUS_INCLUDE_PATH`。\n\n## :rotating_light: 新特性\n\n* 引入了新的 oneDAL 功能：\n  * 添加了对日志记录、跟踪和分析的内部支持。\n  - 支持基于 `tbb::parallel_reduce` 和 `tbb::parallel_deterministic_reduce` 的归约原语。\n  * 通过优化的线性回归，提升了预测和建模性能。\n  * 在所有 x86 CPU 上启用更高位宽的 SIMD 指令，从而提高了向量处理的吞吐量。\n  * 通过距离相关性改进了距离测量的精度。\n  * 提升了 GPU 上大型行数据集的稳定性。\n\n## :beetle: Bug 修复\n\n* 修复了逻辑回归和 Newton-CG 中存在的问题。\n\n## :octocat: 致谢\n\n感谢所有帮助我们实现 2025.6.1 版本发布的人员！\n\n@Alexsandruss, @Alexandr-Solovev, @Vika-F, @david-cortes-intel, @icfaust, @napetrov, @maria-Petrova, @homksei, @ahuber21, @ethanglaser, @razdoburdin, @avolkov-intel\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.5.0...2025.6.0","2025-06-26T14:48:02",{"id":246,"version":247,"summary_zh":248,"released_at":249},144864,"2025.5.0","oneDAL 很高兴地推出 2025.5.0 版本！\n\n## :triangular_flag_on_post: 移除内容与 ABI 兼容性\n\n* 已移除在非 Intel x86 CPU 上分发到 SSE2 代码路径的逻辑\n\n## :rotating_light: 新特性\n\n* 引入了 oneDAL 的新功能：\n  * 为线性回归新增了参数\n  * 针对高维问题引入了非批处理线性回归\n  * 在 CPU 上的决策森林训练算法中增加了超参数支持\n  * 为 `sumWithSIMD` 添加了 SVE 实现\n  * 为 `rocAucScore` 增加了 `float` 数据类型支持\n  * 优化了随机森林算法\n\n## :beetle: 错误修复\n\n* 修复了 PVC 中稀疏 K-means 在数据行数较多时失败的问题\n* 修复了 BFGS 近似可能不正定的问题\n* 为决策树的 `ModelImpl` 类添加了 `isValid()` 方法\n\n## :octocat: 致谢\n\n感谢所有帮助我们实现 2025.5.0 版本发布的人员！\n\n@Alexsandruss、@Alexandr-Solovev、@Vika-F、@david-cortes-intel、@icfaust、@napetrov、@maria-Petrova、@homksei、@ahuber21、@ethanglaser、@razdoburdin、@avolkov-intel\n\n\n\n## 新贡献者\n* @isuruf 在 https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fpull\u002F3026 中完成了首次贡献\n* @richardnorth3 在 https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fpull\u002F3059 中完成了首次贡献\n* @yuejiaointel 在 https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fpull\u002F3108 中完成了首次贡献\n* @theComputeKid 在 https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fpull\u002F3100 中完成了首次贡献\n\n**完整变更日志**：https:\u002F\u002Fgithub.com\u002Fuxlfoundation\u002FoneDAL\u002Fcompare\u002F2025.4.0...2025.5.0","2025-04-23T12:33:30",{"id":251,"version":252,"summary_zh":253,"released_at":254},144868,"2025.0.0","oneDAL is happy to introduce 2025.0.0 release!\r\n\r\n## :triangular_flag_on_post: Removals and ABI Compatibility \r\n\r\n* Starting with the 2025.0 release, the following functionality is removed from oneDAL:\r\n  * openCL kernels\r\n  * DAAL GPU functionality\r\n* Starting with version 2025.0, oneDAL has transitioned from using MKLFPK to oneMKL. For users who wish to build oneDAL manually, it is now possible to use a custom oneMKL build (version 2025.0 or later). This enables greater flexibility and compatibility with recent improvements in oneMKL.\r\n* ABI compatibility is broken as part of the 2025.0 release of oneDAL. The library’s major version is incremented to three to enforce the relinking of existing applications.\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * Enabled calculation of SHAP values for binary classification models required for explainability RF algorithms\r\n  * oneDAL ICX build support\r\n* Improved oneDAL performance for the following algorithms:\r\n  * Decision Forest\r\n\r\n## :beetle: Bug Fixes\r\n\r\n* Fix online SPMD algorithms finalize call\r\n* Addressed `k-Means` timeouts issue on GPU for sparse inputs\r\n\r\n## :x: Deprecation Notice\r\n\r\n* ABI compatibility is to be broken as part of the 2026.0 release of oneDAL. The library’s major version is to be incremented to four to enforce the relinking of existing applications \r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2025.0.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @maria-Petrova, @aepanchi, @emmwalsh, @icfaust, @inteldimitrius, @samir-nasibli, @md-shafiul-alam, @david-cortes-intel\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.7.0...2025.0.0","2024-11-04T13:30:38",{"id":256,"version":257,"summary_zh":258,"released_at":259},144869,"2024.7.0","oneDAL is happy to introduce 2024.7.0 release!\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * k-Means `Init` Sparsity Support\r\n  * SPMD interfaces support: online Linear Regression\r\n  * `2c_mom` reference implementation\r\n\r\n## :beetle: Bug Fixes\r\n\r\n* Correct multiline command handling in `setup_samples.cmake`\r\n* Added missing ONEDAL_EXPORT for sparse Logistic Regression\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.7.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @maria-Petrova, @aepanchi, @emmwalsh, @icfaust, @inteldimitrius, @samir-nasibli, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.6.0...2024.7.0","2024-09-18T13:12:56",{"id":261,"version":262,"summary_zh":263,"released_at":264},144870,"2024.6.0","oneDAL is happy to introduce 2024.6.0 release!\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * Ridge Regression support in oneAPI interfaces\r\n  * Sparsity support for Logistic Regression algorithm\r\n  * Added TBB scheduler handle\r\n\r\n## :beetle: Bug Fixes\r\n\r\n* Fix `loop not vectorized` warning when running with newer icpx compiler\r\n* Fix usage of sequential execution branches\r\n* Fix dpc compilation for OS=win and COMPILER=vc\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.6.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @maria-Petrova, @aepanchi, @emmwalsh, @icfaust, @inteldimitrius, @samir-nasibli, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.5.0...2024.6.0","2024-08-13T15:37:50",{"id":266,"version":267,"summary_zh":268,"released_at":269},144871,"2024.5.0","oneDAL is happy to introduce 2024.5.0 release!\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * DPC++ sparse gemm and gemv primitives\r\n  * Data type support in `read` function\r\n  * Online distributed PCA\r\n  * Sparsity support for `logloss` function primitive\r\n\r\n## :beetle: Bug Fixes\r\n\r\n* sycl `barrier()` deprecation alternative\r\n* Distributed PCA singular values computation fix\r\n* Fix event dependencies for PCA infer on GPU\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.5.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @maria-Petrova, @aepanchi, @emmwalsh, @icfaust, @inteldimitrius, @samir-nasibli, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.4.0...2024.5.0","2024-07-02T10:40:02",{"id":271,"version":272,"summary_zh":273,"released_at":274},144872,"2024.4.0","# oneDAL 2024.4.0\r\noneDAL is happy to introduce 2024.4.0 release!\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * Distributed dbscan support without `allgatherv` (memory saving implementation)\r\n  * Added `global_context` to store information about CPU environment\r\n  * Online distributed basic statistics\r\n  * Hyperparameters added into decision forest classification inference algorithm\r\n  * Added `assume_centered` flag to Covariance\r\n  * Added `dtype` dispatcher for Dataframe Intercahnge Protocol in Python\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.4.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @maria-Petrova, @aepanchi, @emmwalsh, @icfaust, @inteldimitrius, @samir-nasibli, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.3.0...2024.4.0","2024-05-16T14:16:20",{"id":276,"version":277,"summary_zh":278,"released_at":279},144873,"2024.3.0","# oneDAL 2024.3.0\r\noneDAL is happy to introduce 2024.3.0 release!\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * k-Means sparsity support\r\n  * Online support of distributed Covariance algorithm\r\n  * Added `inner_iterations_count` result option to LogisticRegression\r\n\r\n## :beetle: Bug Fixes\r\n\r\n* Fix FPK exports on Windows\r\n* Fix `csr_accessor` crash when moving data from USM or host memory to host memory\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.3.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @aepanchi, @icfaust, @inteldimitrius, @maria-Petrova, @samir-nasibli, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.2.0...2024.3.0","2024-04-11T16:12:05",{"id":281,"version":282,"summary_zh":283,"released_at":284},144874,"2024.2.0","# oneDAL 2024.2.0\r\noneDAL is happy to introduce 2024.2.0 release!\r\n\r\n## :rotating_light: What's New\r\n\r\n* Introduced new oneDAL functionality:\r\n  * Optional results in k-Means inference\r\n  * GPU support of SVD method of PCA algorithm\r\n\r\n* Improved oneDAL performance for the following algorithms:\r\n  * k-Means\r\n\r\n## :beetle: Bug Fixes\r\n\r\n* Fix FPK exports on Windows\r\n* Replace deprecated `sycl::abs` to `sycl::fabs`\r\n\r\n## :octocat: Acknowledgements\r\nThanks to everyone who helped us make 2024.2.0 release possible! \r\n\r\n@avolkov-intel, @Alexsandruss, @ahuber21, @Alexandr-Solovev, @Vika-F, @razdoburdin, @ethanglaser, @napetrov, @aepanchi, @icfaust, @inteldimitrius, @maria-Petrova, @samir-nasibli, @md-shafiul-alam\r\n\r\n**Full Changelog**: https:\u002F\u002Fgithub.com\u002Foneapi-src\u002FoneDAL\u002Fcompare\u002F2024.1.0...2024.2.0","2024-04-02T19:30:21"]