initial-data.json

{"projects":[{"id":"omni","name":"vLLM-Omni","category":"multimodal","url":"https://github.com/vllm-project/vllm-omni","tagline":{"en":"Easy, fast, and cheap omni-modality model serving for everyone.","zh":"让全模态模型服务变得简单、快速、经济，面向每一位开发者。"},"detail":{"en":"Explore the project’s approach to omni-modality model serving. Read the architecture, supported models, and setup instructions in the official repository.","zh":"探索全模态模型服务的开放实现，在项目仓库中了解架构、支持的模型与部署方式。"}},{"id":"afd","name":"afd-plugin","category":"infrastructure","url":"https://github.com/vllm-project/afd-plugin","tagline":{"en":"A vLLM external plugin for Attention-FFN Disaggregation (AFD).","zh":"面向 Attention-FFN 分离（AFD）的 vLLM 外部插件。"},"detail":{"en":"An experimental project exploring Attention-FFN Disaggregation. Consult its repository for current status, supported configurations, and implementation details.","zh":"探索 Attention-FFN 分离的实验性项目。请在仓库中查看当前状态、支持的配置与实现细节。"}},{"id":"router","name":"vLLM Router","category":"infrastructure","url":"https://github.com/vllm-project/router","tagline":{"en":"A high-performance, lightweight request forwarding system for large-scale vLLM deployments.","zh":"面向大规模 vLLM 部署的高性能、轻量级请求转发系统。"},"detail":{"en":"Explore request forwarding for large-scale vLLM deployments, with project documentation and architecture available in the official repository.","zh":"探索大规模 vLLM 部署中的请求转发。在项目仓库中查看架构、文档与使用方式。"}},{"id":"agent","name":"agent-infer","category":"agentic","url":"https://github.com/openJiuwen-ai/agent-infer","tagline":{"en":"A vLLM-based inference framework for agentic workloads.","zh":"基于 vLLM，面向智能体工作负载的推理框架。"},"detail":{"en":"Discover an inference framework focused on agentic workloads. Explore the framework architecture and follow the repository for current setup and integration guidance.","zh":"聚焦智能体工作负载，探索推理框架架构，并在项目仓库中查看最新的配置与集成指南。"}},{"id":"rlt","name":"vllm-rlt","category":"recurrent","url":"https://github.com/ThinkFlowLab/vllm-rlt","tagline":{"en":"A standalone inference and serving engine for recurrent looped transformers.","zh":"面向循环迭代 Transformer（recurrent looped transformers）的独立推理与服务引擎。"},"detail":{"en":"Explore standalone inference and serving for recurrent looped transformers. The poster-matched diagram illustrates the repeated-pass concept; consult the repository for implementation details.","zh":"探索循环迭代 Transformer 的独立推理与模型服务。海报同款架构图展示了多轮计算的概念；具体实现请参阅项目仓库。"}},{"id":"mlx","name":"vllm-omni-mlx","category":"multimodal","url":"https://github.com/ThinkFlowLab/vllm-omni-mlx","tagline":{"en":"A lightweight Apple Silicon counterpart of vLLM-Omni.","zh":"面向 Apple Silicon 的轻量级 vLLM-Omni 对应实现。"},"detail":{"en":"Explore a lightweight approach to omni-modality inference on Apple Silicon. Check the repository for supported features, model compatibility, and local setup.","zh":"探索 Apple Silicon 上的轻量级全模态推理。在项目仓库中了解支持的功能、模型兼容性与本地使用方式。"}},{"id":"inferrsi","name":"InferRSI","category":"optimization","url":"https://github.com/InferMatrixLab/InferRSI","tagline":{"zh":"以可验证的实验，推动推理优化与专用引擎探索。","en":"Turn inference bottlenecks into verifiable improvements."},"detail":{"zh":"面向编码智能体的推理优化 CLI 工作台：从性能剖析、重复 A/B 实验与质量检查，到可追溯的证据报告和经验复用。用户提供实验命令并驱动代码修改；并非已实现的全自动优化系统。下方 InferRSI 专区展示两个专用语音引擎案例。","en":"A CLI workbench for inference optimization with coding agents: profiling, repeated A/B experiments, quality gates, traceable evidence and reusable lessons. Users supply experiment commands and drive code changes; this is not a fully autonomous optimizer. Explore two dedicated speech-engine examples in the InferRSI section below."}}],"copy":{"en":{"filterOptimization":"Optimization","navVision":"Vision","navProjects":"Projects","navExplore":"Explore","navBuild":"Build with us","heroEyebrow":"OPEN SOURCE. OPEN POSSIBILITIES.","heroDescription":"An open-source constellation for the next era of AI inference.","heroSubline":"让每一次推理，都成为下一次进化的起点。","heroCta":"Explore the ecosystem","heroSecondary":"See the connections","scroll":"SCROLL TO DISCOVER","heroBottom":"SEVEN PROJECTS. SHARED INNOVATION.","motion":"MOTION ON","motionOff":"MOTION OFF","visionKicker":"THE VISION","visionTitle":"Intelligence moves forward.<br><span>Infrastructure should, too.</span>","visionDescription":"New modalities. New architectures. New ways to think. The future of inference is being built in the open.","visionSecondary":"InferMatrix brings seven open-source projects into view, connecting ideas across models, hardware, and serving infrastructure.","visionCta":"Meet the ecosystem","projectsKicker":"THE OPEN-SOURCE ECOSYSTEM","projectsTitle":"Seven projects.<br><span>One open horizon.</span>","projectsDescription":"Explore the code.<br>Understand the architecture.<br>Find your next building block.","filterAll":"All projects","filterMultimodal":"Multimodal","filterInfrastructure":"Infrastructure","filterAgentic":"Agentic","filterRecurrent":"Recurrent","directionOptimization":"Autonomous inference optimization","projectFootnote":"Open-source projects from across the community. Follow each repository for documentation, licensing, and contribution guidelines.","exploreKicker":"CONNECTED POSSIBILITIES","exploreTitle":"Find your path<br><span>through the matrix.</span>","exploreDescription":"Different challenges. Shared curiosity.<br>Select a direction to explore the ecosystem.","conceptual":"CONCEPTUAL ECOSYSTEM MAP","exploreRelated":"Explore related projects","conceptNote":"An illustrative map of focus areas, not a deployment topology or compatibility guarantee.","sourceKicker":"THE NEXT CHAPTER IS OPEN","sourceTitle":"Build what<br><span>comes next.</span>","sourceDescription":"Start with a repository. Follow an idea.<br>Move inference forward, together.","sourceCta":"Find your project","posterCta":"View the brand poster","backTop":"Back to top","footerNote":"An open-source ecosystem showcase.","downloadPoster":"Download poster ↗","posterTitle":"Six projects. Shared innovation.","posterNote":"Original supplied artwork. The website uses the updated vllm-rlt description: recurrent looped transformers.","download":"Download ↗","details":"Architecture & details","github":"GitHub","preview":"VIEW ARCHITECTURE","viewRepo":"View repository","architectureOfficial":"Official architecture image","architectureDocsVector":"Vector diagram redrawn from official documentation","architectureVector":"Vector diagram redrawn from official architecture","architecturePosterVector":"Vector diagram redrawn from the supplied poster","architectureSource":"View source","architectureAlt":"architecture diagram","architectureUnavailable":"Architecture image is being prepared.","zoomIn":"Zoom in","zoomOut":"Zoom out","zoomReset":"Fit to window","zoomActual":"Original size","zoomControls":"Architecture image zoom controls","zoomLevel":"Zoom","zoomHint":"Use + / − to zoom. Scroll to explore the diagram.","viewerLabel":"Scrollable architecture image","openOriginal":"Open original","openVector":"Open vector","downloadArchitecture":"Download image","showing":"SHOWING","direction":"DIRECTION","closeDetails":"Close project details","closePoster":"Close poster","openNav":"Open navigation","closeNav":"Close navigation","pauseMotion":"Pause decorative motion","resumeMotion":"Resume decorative motion","switchLang":"Switch language to Chinese","skip":"Skip to content","mapLabel":"Conceptual map highlighting {names} around an inference core"},"zh":{"filterOptimization":"优化探索","navVision":"品牌愿景","navProjects":"开源项目","navExplore":"探索生态","navBuild":"一起构建","heroEyebrow":"开放源码，无限可能。","heroDescription":"为 AI 推理的下一个时代，连接开放创新的力量。","heroSubline":"让每一次推理，都成为下一次进化的起点。","heroCta":"探索开源生态","heroSecondary":"发现技术连接","scroll":"向下探索","heroBottom":"七个开源项目，共享创新。","motion":"动态效果 开","motionOff":"动态效果 关","visionKicker":"品牌愿景","visionTitle":"智能不断向前。<br><span>基础设施，也应持续进化。</span>","visionDescription":"新的模态，新的架构，新的思考方式。推理的未来，正在开放协作中构建。","visionSecondary":"InferMatrix 汇聚七个开源项目，连接模型、硬件与推理服务基础设施中的创新思路。","visionCta":"认识开源生态","projectsKicker":"开源项目生态","projectsTitle":"七个项目。<br><span>共同打开新的可能。</span>","projectsDescription":"探索代码，理解架构。<br>找到你的下一个构建起点。","filterAll":"全部项目","filterMultimodal":"多模态","filterInfrastructure":"基础设施","filterAgentic":"智能体","filterRecurrent":"循环推理","directionOptimization":"推理自主优化","projectFootnote":"这里展示来自开源社区的项目。文档、许可证与贡献指南，请以各项目仓库为准。","exploreKicker":"连接更多可能","exploreTitle":"在技术矩阵中，<br><span>找到你的探索方向。</span>","exploreDescription":"不同的挑战，共同的好奇心。<br>选择一个方向，探索相关项目。","conceptual":"生态方向概念图","exploreRelated":"查看相关项目","conceptNote":"此图展示项目的研究与应用方向，不代表部署拓扑或兼容性保证。","sourceKicker":"下一篇章，由开放开启","sourceTitle":"一起构建，<br><span>下一个可能。</span>","sourceDescription":"从一个仓库开始，让一个想法生长。<br>一起推动推理技术向前。","sourceCta":"找到你的项目","posterCta":"查看品牌海报","backTop":"返回顶部","footerNote":"开放推理生态展示。","downloadPoster":"下载品牌海报 ↗","posterTitle":"六个开源项目，共享创新。","posterNote":"此处保留原始品牌海报。网站已采用更新后的 vllm-rlt 描述：面向 recurrent looped transformers。","download":"下载海报 ↗","details":"架构与详情","github":"GitHub","preview":"查看架构","viewRepo":"前往项目仓库","architectureOfficial":"官方架构原图","architectureDocsVector":"依据官方文档重绘（矢量）","architectureVector":"依据官方架构重绘（矢量）","architecturePosterVector":"依据原始海报重绘（矢量）","architectureSource":"查看来源","architectureAlt":"架构图","architectureUnavailable":"架构图正在准备中。","zoomIn":"放大","zoomOut":"缩小","zoomReset":"适应窗口","zoomActual":"原始尺寸","zoomControls":"架构图缩放控制","zoomLevel":"缩放","zoomHint":"点击 + / − 缩放，滚动查看架构细节。","viewerLabel":"可滚动的架构图","openOriginal":"打开原图","openVector":"打开矢量图","downloadArchitecture":"下载架构图","showing":"显示","direction":"探索方向","closeDetails":"关闭项目详情","closePoster":"关闭品牌海报","openNav":"打开导航","closeNav":"关闭导航","pauseMotion":"暂停装饰动画","resumeMotion":"开启装饰动画","switchLang":"Switch language to English","skip":"跳转到主要内容","mapLabel":"概念图：围绕推理核心，突出展示 {names}"}},"architectureAssets":{"inferrsi":{"src":"assets/architecture/inferrsi.svg","width":1800,"height":1120,"kind":"docs-vector","sourceUrl":"https://github.com/InferMatrixLab/InferRSI/blob/main/docs/assets/modular-components-en.svg"},"omni":{"src":"assets/architecture/omni.png","width":1757,"height":895,"kind":"official","sourceUrl":"https://github.com/vllm-project/vllm-omni/blob/main/docs/source/architecture/omni-modality-model-architecture.png"},"afd":{"src":"assets/architecture/afd.svg","width":1280,"height":720,"kind":"official","sourceUrl":"https://github.com/vllm-project/afd-plugin/blob/main/docs/assets/vllm-afd-plugin-architecture.svg"},"agent":{"src":"assets/architecture/agent-poster.svg","width":1295,"height":1260,"kind":"poster-vector","sourceUrl":"assets/infermatrix-poster.png"},"rlt":{"src":"assets/architecture/rlt-poster.svg","width":1544,"height":1028,"kind":"poster-vector","sourceUrl":"assets/infermatrix-poster.png"},"mlx":{"src":"assets/architecture/mlx-poster.svg","width":1792,"height":1044,"kind":"poster-vector","sourceUrl":"assets/infermatrix-poster.png"},"router":{"src":"assets/architecture/router-poster.svg","width":1800,"height":1360,"kind":"poster-vector","sourceUrl":"assets/infermatrix-poster.png"}}}