diff --git a/README.md b/README.md
index bde21af..5f24fac 100644
--- a/README.md
+++ b/README.md
@@ -14,7 +14,7 @@ A bilingual, buildless website showcasing seven open-source inference projects:
python3 -m http.server 8000 --bind 127.0.0.1
```
-Open [http://localhost:8000](http://localhost:8000). The deployable files are already at the repository root: `index.html`, `styles.css`, `assistant.css`, `app.js`, `showcase.js`, `inferrsi.js`, `assistant.js`, `video-resources.js`, `robots.txt`, `sitemap.xml`, and `assets/`.
+Open [http://localhost:8000](http://localhost:8000). The deployable files are already at the repository root: `index.html`, `styles.css`, `assistant.css`, `app.js`, `showcase.js`, `inferrsi.js`, `assistant.js`, `resource-library.js`, `robots.txt`, `sitemap.xml`, and `assets/`.
The stylesheet requests DM Sans and Manrope from Google Fonts. System-font fallbacks remain available when the external font service cannot be reached.
@@ -26,10 +26,11 @@ Optional maintenance checks use Node.js 18+ built-in modules only:
```sh
node --check app.js
-node --check video-resources.js
+node --check resource-library.js
node --check prepare-initial-content.cjs
node --check qa-interactions.cjs
node qa-interactions.cjs
+node qa-resources.cjs
```
The interaction suite reads the current `index.html` and runs `app.js` in a simulated DOM using Node's VM. It covers filtering, bilingual content, keyboard navigation, project/poster dialogs, focus return, motion preferences, blocked storage, and all six architecture viewers. This is a logic regression check; it does not replace real-browser visual or accessibility testing.
@@ -57,6 +58,8 @@ Bind each custom domain in its Pages project. The registrar must delegate `infer
## Latest additions
+- A bilingual Articles column with 14 verified technical articles and practical guides covering all seven projects, separate filters and search, and six entries per page.
+
- A bilingual Video resources column for external learning tutorials and project talks, with project/type filters and search. Eleven links cover vLLM-Omni, afd-plugin and vLLM Router, with available publisher information and verification sources retained.
- A repo-grouped video showcase with five vLLM-Omni videos and two InferRSI English demos, local poster previews, click-to-play and fullscreen controls.
- InferRSI as a seventh project and an evidence-driven optimization direction, with a brand-styled, scalable architecture SVG.
@@ -81,7 +84,7 @@ for the optional dependency-free Node preview and interaction checklist.
## 视频资源 / Video resources
-首页导航的「视频资源」指向 `#video-resources`,与「演示展示」分别维护。专栏收录学习教程、技术分享和宣讲回放的外部链接,点击后在新窗口打开原发布平台,不嵌入或下载视频。
+首页导航的「视频专栏」指向 `#video-resources`,与「演示展示」分别维护。专栏收录学习教程、技术分享和宣讲回放的外部链接,点击后在新窗口打开原发布平台,不嵌入或下载视频。
资源清单集中在 `assets/video-resources.json`。当前收录 11 条:vLLM-Omni 8 条、afd-plugin 2 条、vLLM Router 1 条;其他四个项目暂未找到可核实的教程或宣讲回放。用户指定合集分集时保留 `p` 参数,并移除分享链接中的跟踪参数;可用 B 站官方分页元数据核实分集标题、讲者(标题明确标注时)与时长。每添加一条资源,填写以下格式;示例仅说明字段,不是已收录的视频:
@@ -105,10 +108,20 @@ for the optional dependency-free Node preview and interaction checklist.
每条已收录资源还保留 `sourceUrl`、`verifiedAt` 和 `sourceNote`,供维护者核对来源、核实日期及证据范围;有依据时记录 `eventDate`。这些维护字段不显示在视频卡片上。`publishedAt` 只填写已确认的视频上传日期,不能用会议日期或配套文章日期代替。中英文标题是便于浏览的编辑文案;原始标题与信息请以来源页面为准。首批资料来源包括 vLLM 官方 Office Hours、Conf42、Red Hat Developer、B 站原视频页面以及 vLLM 小课堂的原视频与课件索引。B 站的部分页面/API 会对自动抓取返回 412;其链接由原始页面索引或配套资料交叉核对,未逐条验证实际播放。
-The library is rendered by `video-resources.js`. Project counts reflect the full catalog; results combine project, content-type, and text filters. Results are paginated at 6 videos per page, with bilingual previous/next controls and the current visible range. Changing a filter or search resets to page one; changing language preserves the current page and filters. Pagination is hidden for a single page, empty results, loading and errors. Search covers both languages, project names, platforms, publishers, and speakers. Empty catalogs, no matches, and failed requests have distinct states; failed requests can be retried.
+The library is rendered by `resource-library.js`. Project counts reflect the full catalog; results combine project, content-type, and text filters. Results are paginated at 6 videos per page, with bilingual previous/next controls and the current visible range. Changing a filter or search resets to page one; changing language preserves the current page and filters. Pagination is hidden for a single page, empty results, loading and errors. Search covers both languages, project names, platforms, publishers, and speakers. Empty catalogs, no matches, and failed requests have distinct states; failed requests can be retried.
The catalog is fetched with `cache: 'no-store'` and a fresh request parameter so edits to the JSON appear on the next page reload, even when the document or previous JSON response was cached. Changing only the page URL's query string does not invalidate a separately fetched resource. The script entry also carries a version parameter to refresh older copies of the loader.
+## 文章专栏 / Articles
+
+首页导航的「文章专栏」指向 `#article-resources`。首批收录 14 篇官方博客、社区技术解读与项目实践指南,覆盖七个项目。点击「阅读原文」前往原发布平台;中英文标题和简介为本站编辑文案,正文以原文为准。
+
+文章清单位于 `assets/article-resources.json`,与视频分开维护。字段沿用视频资源结构;类型为 `tutorial`(学习教程)或 `technical`(技术解读),可选作者字段为 `author`,无需填写 `speaker` 或 `duration`。每条记录保留 `sourceUrl`、`sourceNote` 和 `verifiedAt`,核实日期使用 `YYYY-MM-DD`。只有来源明确提供发布日期时才填写 `publishedAt`;持续更新的项目文档不使用仓库修改时间代替发布日期。
+
+视频与文章在同一个 `#learning-resources` 区域,通过「视频专栏 / 文章专栏」页签切换,一次显示一个栏目。普通访问默认选择视频;`#video-resources` 与 `#article-resources` 深链接、首页对应导航仍可直达指定栏目。页签支持左右方向键循环切换及 Home / End;无 JavaScript 时保留两个资源清单的访问入口。
+
+`resource-library.js` initializes the video and article libraries independently using the same rendering and filtering logic. Each keeps its own project/type filters, search, page, loading and retry state across tab and language changes. Both show at most six entries per page. Article search includes authors; links and text use the same HTTPS validation and escaping as videos. `qa-resources.cjs` exercises both libraries together, including accessible tab navigation, deep links, isolated fetch failures, invalid entries, pagination and language changes.
+
## 素材与许可 / Assets and licensing
See [ATTRIBUTION.md](ATTRIBUTION.md) for asset-specific sources and [licenses/](licenses/) for the retained upstream licenses. The two official upstream diagrams retain their Apache-2.0 terms. The logo, poster, and poster-based vector redraws are not assigned an open-source license by this repository. No blanket license is asserted for all website code and branding.
diff --git a/SHA256SUMS b/SHA256SUMS
index 7c2e72d..269e250 100644
--- a/SHA256SUMS
+++ b/SHA256SUMS
@@ -5,7 +5,7 @@ dc2b68dea37313b64e7a5e585f68be9164e3ebfae84ed5614a28519159a68454 ATTRIBUTION.md
fc44df75947d9534405fc9f5401de083dd2fab3e474a5f781348137ad306e10c BENCHMARK_EVIDENCE.json
00fe86be6eb9b9d39857f1833be8ab6d9ae6d39a9117a01ccfcdf3dc46a9f956 INFERRSI.md
9b37f8f61e35ca9998499716bd299a03f05d81208acae3bb35900807a0fc55ad LOCAL_PREVIEW.md
-0c48cee1c5e4e1f0b193f87fa49e1b8647ba876cb422360273569225f5b74c4e README.md
+cd924672b113b8d4b8cdc236a372beaf20641e2a04450a72a94db4eab511122a README.md
5655fa3532ac07d339e7c44de9118a08d2a83bf1cfeadef7395f08e440480902 SHOWCASE.md
0bc3d5f7bd34ac53892f39ab97de48a9d771ddbfc35676255fa46292e4dcae81 app.js
53cf5937ee480102fbabd8e26a52adf1e936a49c2c24d0fc1948465982466017 assets/architecture/afd.svg
@@ -20,6 +20,7 @@ ae0a52e430db46ca989b63eb679b13db0a6b90516fc70f3269172bd1c5fc2ff6 assets/archite
2972f105042a61cb960d2923dd2916d3c6e0644f27cfe73f55fedf42eeed781b assets/architecture/rlt.png
5c35d3277b107940d1cf501b5344d5e3a06f68f60cda6b34dfa274b0defdd731 assets/architecture/router-poster.svg
cc3545981d0ec428efe6dd61d35901b0fc3f8beb50f696616f5eb30a1689faae assets/architecture/router.svg
+94643191d8af94adca001ac35dac5766c065805fed1c911fdd195bd9e5b608ab assets/article-resources.json
1e43c9c8b4dc78086790a41fd46369141802e12e592349f08a804107a35fa1a3 assets/favicon.svg
7b1b2d9403e7a8b83cd695adc6d5a7c9ff5b697225de3149057c0f708c1fddc5 assets/infermatrix-logo.png
79d4c4265b2d7c5635056ab49064469ab48c98dfb326e2e1fcbfbb1c3effe5ad assets/infermatrix-poster.png
@@ -46,7 +47,7 @@ f1e5d4f2cc41c9c1da69dfb3523c5659012ac8169d06c267b165af5ecd2aa698 assets/showcas
30e9b46a6d06ab864bbc7ca0289143822d1192392074dc965b0683559f6911e0 assistant.css
71ba38902b5a33b28d7bb66f5a90d83a73cbc4ca9b0d037289b7abce221a30c2 assistant.js
1c13002720c8d65a3fc98084c708b9058827ab66740ae52105991cfebc60eb8e dom-fixture.json
-cc9f9f5815ecfb81c9df0ea4de1f40088fdd0ab17e336502c5cd21d3fa73e413 index.html
+14f43be31694e4e0a5b416917ee02fa826a225fb13386ea6ecc8a1602474d878 index.html
9ec768282a4d526da0c76048ae1e037db9103084bd8e0d5662b22930f433660d inferrsi.js
ae0fe01b8c81c20e4a607f8608b84ac0df3b42f9e38bcaaefde92cda5c63a782 initial-data.json
0e8566ea362276878e1f38f97da49964dbea301c15d4aa9d727f97a4c39fd943 licenses/afd-plugin-LICENSE.txt
@@ -54,8 +55,9 @@ c71d239df91726fc519c6eb72d318ec65820627232b2f796219e87dcf35d0ab4 licenses/vllm-
19cc4e8b177647501dc4c8fa02ce38e90f0c1f3efe70fe2aa3409930bde7aae1 local-preview.cjs
250f9ee919523549bbabe67fa4585a2772a405f8364b2a370b4020e61aedb8b6 prepare-initial-content.cjs
5462843e0ed8edf7fbc2fcdca220ff496cbd0533a914ddd9c91d1e26b40d8f08 qa-interactions.cjs
+587d81c6a85f504e6df43ce8ae59b70aec771b521e1bdafa15e6c0d63e25619c qa-resources.cjs
+071cfc59cbb253dbaec6a544db376be8c466c35ab6590ea9302e151cf4850b09 resource-library.js
51a3881b90c67476605580f3a859becf174b8bf67d37c5c1200324613f2ab42d robots.txt
79c5f0e08b8dc1cfcae28c9e95851c2ee4f3ab7e7eb9dce09ee5303522e17d22 showcase.js
741e991c7d0bc12c49acc061ba697dcc5e57946380c708cc823f7fadcb3d6485 sitemap.xml
-1e7d9d7df6e610ec11f441989b64bc51fed2770d674f5f1b6a04f718240e259d styles.css
-cc80a9dffa71fd10669b5385dfa0771f9f64052d8e1bb9bc6b311bdadf0a77ad video-resources.js
+71599431c0c1bc0bc73f7105ac319bb1ba6393e4532189aaad524729477d14e5 styles.css
diff --git a/assets/article-resources.json b/assets/article-resources.json
new file mode 100644
index 0000000..5b5fd59
--- /dev/null
+++ b/assets/article-resources.json
@@ -0,0 +1,275 @@
+[
+ {
+ "id": "omni-announcement",
+ "project": "omni",
+ "type": "technical",
+ "title": {
+ "zh": "vLLM-Omni 发布:简单、高效的全模态模型服务",
+ "en": "Announcing vLLM-Omni: Easy, Fast, and Cheap Omni-Modality Model Serving"
+ },
+ "description": {
+ "zh": "官方介绍全模态服务的目标、异构模型流水线和分阶段资源配置,并提供入门入口。",
+ "en": "The official introduction to omni-modal serving, heterogeneous model pipelines and stage-level resource allocation, with links to get started."
+ },
+ "url": "https://vllm.ai/blog/2025-11-30-vllm-omni",
+ "platform": "vLLM Blog",
+ "author": "vLLM-Omni Team",
+ "publishedAt": "2025-11-30",
+ "sourceUrl": "https://vllm.ai/blog/2025-11-30-vllm-omni",
+ "sourceNote": "Title, byline, publication date and technical scope verified on the original official article. Covers motivation, architecture, key features and getting started; roadmap items are not asserted as completed features.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "omni-redhat-qwen3-architecture",
+ "project": "omni",
+ "type": "technical",
+ "title": {
+ "zh": "走进 vLLM-Omni 架构:服务 Qwen3-Omni",
+ "en": "Inside the vLLM-Omni Architecture: Serving Qwen3-Omni"
+ },
+ "description": {
+ "zh": "通过保险理赔演示理解 Thinker、Talker、Code2Wav 三阶段服务、OmniConnector 和流式流水线。",
+ "en": "An insurance-claim demo explains the Thinker, Talker and Code2Wav stages, OmniConnector transport and streaming pipeline execution."
+ },
+ "url": "https://developers.redhat.com/articles/2026/07/01/inside-vllm-omni-architecture-serving-qwen3-omni",
+ "platform": "Red Hat Developer",
+ "author": "Isaac Tigges",
+ "publishedAt": "2026-07-01",
+ "sourceUrl": "https://developers.redhat.com/articles/2026/07/01/inside-vllm-omni-architecture-serving-qwen3-omni",
+ "sourceNote": "Title, byline, publication date and technical scope verified on the original Red Hat Developer article. This is a complete written architecture walkthrough with a companion video; the article date and byline describe the article.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "omni-qwen3-multistage-optimization",
+ "project": "omni",
+ "type": "technical",
+ "title": {
+ "zh": "Qwen3-Omni 多阶段服务:实践经验与性能优化",
+ "en": "Experience and Lessons Learned from Serving Multi-Stage Qwen3-Omni in vLLM-Omni"
+ },
+ "description": {
+ "zh": "逐阶段讲解批处理、CUDA Graph、异步分块、异步输出和副本扩容,并提供受控测试结果。",
+ "en": "A stage-by-stage walkthrough of batching, CUDA Graphs, async chunking, async output and replica scaling, supported by controlled benchmarks."
+ },
+ "url": "https://vllm.ai/blog/2026-07-01-qwen3-omni-optimization",
+ "platform": "vLLM Blog",
+ "author": "vLLM-Omni Team and Ant Group SCT Team",
+ "publishedAt": "2026-07-01",
+ "sourceUrl": "https://vllm.ai/blog/2026-07-01-qwen3-omni-optimization",
+ "sourceNote": "Title, byline, publication date and optimization chapters verified on the original official article. The summary describes mechanisms and controlled validation without generalizing measured speedups across configurations.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "omni-official-quickstart",
+ "project": "omni",
+ "type": "tutorial",
+ "title": {
+ "zh": "vLLM-Omni 快速入门:离线推理与在线服务",
+ "en": "vLLM-Omni Quickstart: Offline Inference and Online Serving"
+ },
+ "description": {
+ "zh": "从环境与安装开始,运行离线图像生成,并通过 OpenAI 兼容接口启动在线服务。",
+ "en": "Set up the environment, run offline image generation and launch online serving through an OpenAI-compatible API."
+ },
+ "url": "https://docs.vllm.ai/projects/vllm-omni/en/latest/getting_started/quickstart/",
+ "platform": "vLLM-Omni Docs",
+ "originalTitle": "Quickstart",
+ "sourceUrl": "https://docs.vllm.ai/projects/vllm-omni/en/latest/getting_started/quickstart/",
+ "sourceNote": "Topic, installation steps, offline image-generation examples and online-serving examples verified on the project-maintained guide. This is a continuously updated latest document; no publication date or individual author is asserted.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "afd-plugin-announcement",
+ "project": "afd",
+ "type": "technical",
+ "title": {
+ "zh": "vLLM AFD Plugin 发布:分离 Attention 与 FFN",
+ "en": "Announcing vLLM AFD Plugin: Disaggregating Attention and FFN for Flexible MoE Serving"
+ },
+ "description": {
+ "zh": "介绍实验性 AFD 插件的 Attention / FFN 服务划分、Connector 接口、GPU / NPU 支持与部署方式。",
+ "en": "An introduction to the experimental AFD plugin, its Attention and FFN services, connector interface, GPU/NPU support and deployment recipes."
+ },
+ "url": "https://vllm.ai/blog/2026-07-23-vllm-afd-plugin",
+ "platform": "vLLM Blog",
+ "author": "AFD Plugin Contributors",
+ "publishedAt": "2026-07-23",
+ "sourceUrl": "https://vllm.ai/blog/2026-07-23-vllm-afd-plugin",
+ "sourceNote": "Title, byline, publication date, architecture and deployment sections verified on the original official article. The article explicitly calls the plugin experimental and limits support claims to validated configurations.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "afd-fastafd-technical-walkthrough",
+ "project": "afd",
+ "type": "technical",
+ "title": {
+ "zh": "大模型推理的内存与计算解耦:AFD 与 FastAFD",
+ "en": "Memory–Compute Disaggregation with AFD and FastAFD"
+ },
+ "description": {
+ "zh": "社区技术解读,拆解 Attention / FFN 分离、服务角色、双微批流水线和 FastAFD 优化及其适用边界。",
+ "en": "A community walkthrough of Attention/FFN disaggregation, runtime roles, dual-microbatch pipelines and FastAFD optimizations and their limits."
+ },
+ "url": "https://princepride.github.io/live-streaming-tools/fastafd_vllm_plugin/final/blog/",
+ "platform": "vLLM 小课堂技术博客",
+ "originalTitle": "大模型推理的内存与计算解耦:AFD 架构解析与性能验证",
+ "sourceUrl": "https://princepride.github.io/live-streaming-tools/fastafd_vllm_plugin/final/blog/",
+ "sourceNote": "Verified as a complete community technical article with companion lesson materials. The display title follows the article's navigation title; originalTitle retains its H1. No individual author or publication date is asserted, and it is not labeled as a project-official announcement.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "router-official-announcement",
+ "project": "router",
+ "type": "technical",
+ "title": {
+ "zh": "vLLM Router:面向大规模服务的高性能 P/D 感知负载均衡",
+ "en": "vLLM Router: A High-Performance and Prefill/Decode Aware Load Balancer for Large-scale Serving"
+ },
+ "description": {
+ "zh": "官方介绍请求路由、会话亲和、Prefill / Decode 分离,以及服务发现、容错和监控能力。",
+ "en": "The official introduction to request routing, session affinity, prefill/decode disaggregation, service discovery, reliability and observability."
+ },
+ "url": "https://vllm.ai/blog/2025-12-13-vllm-router-release",
+ "platform": "vLLM Blog",
+ "author": "vLLM Team",
+ "publishedAt": "2025-12-13",
+ "sourceUrl": "https://vllm.ai/blog/2025-12-13-vllm-router-release",
+ "sourceNote": "Title, byline, publication date and technical scope verified on the original official article. Its repository link identifies vllm-project/router, distinct from vLLM Semantic Router. No benchmark numbers are generalized in the catalog summary.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "router-load-balancing-guide",
+ "project": "router",
+ "type": "tutorial",
+ "title": {
+ "zh": "vLLM Router 负载均衡策略指南",
+ "en": "vLLM Router Load Balancing Policies"
+ },
+ "description": {
+ "zh": "比较一致性哈希、二选一、缓存感知和 KV 感知等策略,学习会话亲和与路由配置。",
+ "en": "Compare consistent hashing, power-of-two, cache-aware and KV-aware policies, with examples for session affinity and routing configuration."
+ },
+ "url": "https://github.com/vllm-project/router/blob/main/docs/load_balancing/README.md",
+ "platform": "vLLM Router Docs",
+ "originalTitle": "Load Balancing Policies",
+ "sourceUrl": "https://github.com/vllm-project/router/blob/main/docs/load_balancing/README.md",
+ "sourceNote": "Verified as the complete project-maintained load-balancing guide, including six routing policies, session-affinity key priorities and configuration examples. This README is a dedicated docs/load_balancing topic, not the repository homepage or a directory listing. No publication date or individual author is asserted.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "omni-diffusion-continuous-batching-pid-article",
+ "project": "omni",
+ "type": "technical",
+ "title": {
+ "zh": "vLLM-Omni:Diffusion 动态批处理与 PiD",
+ "en": "vLLM-Omni: Diffusion Continuous Batching and PiD"
+ },
+ "description": {
+ "zh": "社区技术解读,学习按降噪步调度、动态组批与异构分辨率处理,理解 PiD 的机制及性能边界。",
+ "en": "A community walkthrough of step-level scheduling, dynamic batching, mixed-resolution execution and the mechanisms and limits of PiD."
+ },
+ "url": "https://princepride.github.io/live-streaming-tools/vllm_omni_diffusion_continuous_batching/final/blog/",
+ "platform": "vLLM 小课堂技术博客",
+ "originalTitle": "vLLM-Omni 核心解析:Diffusion 异步批处理与 PiD 架构实践",
+ "sourceUrl": "https://princepride.github.io/live-streaming-tools/vllm_omni_diffusion_continuous_batching/final/blog/",
+ "sourceNote": "Verified as a complete community technical article covering step-level scheduling, dynamic batching, heterogeneous resolutions and PiD, with companion lesson materials. No individual author or publication date is asserted, and it is not labeled as a project-official announcement.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "agentinfer-quick-start",
+ "project": "agent",
+ "type": "tutorial",
+ "title": {
+ "zh": "AgentInfer 快速入门:安装与启用智能体感知调度",
+ "en": "AgentInfer Quick Start: Installation and Agent-Aware Scheduling"
+ },
+ "description": {
+ "zh": "从准备 vLLM 环境、安装 AgentInfer 到启用调度器和验证服务,完成第一次部署。",
+ "en": "Prepare vLLM, install AgentInfer, enable its scheduler, and verify the running service."
+ },
+ "url": "https://github.com/openJiuwen-ai/agent-infer/blob/main/docs/en/tutorial/01-quick-start.md",
+ "platform": "GitHub Docs",
+ "publisher": "openJiuwen / AgentInfer",
+ "sourceUrl": "https://github.com/openJiuwen-ai/agent-infer/blob/main/docs/en/tutorial/01-quick-start.md",
+ "sourceNote": "Verified as the official quick-start guide covering the vLLM environment, installation, --agentinfer startup and models-endpoint validation. The API modification timestamp is not used as a publication date; no individual author is asserted.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "rlt-installation-first-request",
+ "project": "rlt",
+ "type": "tutorial",
+ "title": {
+ "zh": "vllm-rlt 用户指南:从安装到第一次推理请求",
+ "en": "vllm-rlt User Guide: From Installation to the First Request"
+ },
+ "description": {
+ "zh": "按照逐步指南在 NVIDIA GPU 上启动 Ouro-1.4B,发送普通和流式请求,并了解 CLI 与 Python API。",
+ "en": "Serve Ouro-1.4B on an NVIDIA GPU, send regular and streaming requests, and explore the CLI and Python API."
+ },
+ "url": "https://github.com/InferMatrixLab/vllm-rlt/blob/main/docs/launching.md",
+ "platform": "GitHub Docs",
+ "publisher": "InferMatrixLab",
+ "sourceUrl": "https://github.com/InferMatrixLab/vllm-rlt/blob/main/docs/launching.md",
+ "sourceNote": "Verified as the official step-by-step user guide covering prerequisites, model download, serving, generation requests and shutdown. Uses the canonical InferMatrixLab repository after its migration from ThinkFlowLab. The API modification timestamp is not used as a publication date; no individual author is asserted.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "mlx-speech-tts-guide",
+ "project": "mlx",
+ "type": "tutorial",
+ "title": {
+ "zh": "vllm-omni-mlx 语音合成指南:音色、流式输出与声音克隆",
+ "en": "vllm-omni-mlx Speech Guide: Voices, Streaming, and Voice Cloning"
+ },
+ "description": {
+ "zh": "学习 TTS 接口的音色选择、风格指令、声音克隆和流式参数,并了解不同模型的使用差异。",
+ "en": "Learn TTS voice selection, style instructions, cloning, and streaming controls, with differences between model checkpoints."
+ },
+ "url": "https://github.com/InferMatrixLab/vllm-omni-mlx/blob/main/docs/speech.md",
+ "platform": "GitHub Docs",
+ "publisher": "InferMatrixLab",
+ "originalTitle": "Speech (TTS) Usage",
+ "sourceUrl": "https://github.com/InferMatrixLab/vllm-omni-mlx/blob/main/docs/speech.md",
+ "sourceNote": "Verified as the official Speech (TTS) Usage guide with API examples and model-specific differences. Uses the canonical InferMatrixLab repository after its migration from ThinkFlowLab. The API modification timestamp is not used as a publication date; no individual author is asserted.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "inferrsi-quickstart",
+ "project": "inferrsi",
+ "type": "tutorial",
+ "title": {
+ "zh": "InferRSI 快速入门:基准对比、质量验证与证据报告",
+ "en": "InferRSI Quickstart: Benchmarks, Quality Checks, and Evidence Reports"
+ },
+ "description": {
+ "zh": "了解安装与环境检查、基准和质量命令约定,以及如何读取实验报告和导出引用上下文。",
+ "en": "Install and inspect the workbench, define benchmark and quality commands, read experiment reports, and export cited context."
+ },
+ "url": "https://github.com/InferMatrixLab/InferRSI/blob/main/docs/quickstart.md",
+ "platform": "GitHub Docs",
+ "publisher": "InferMatrixLab",
+ "originalTitle": "Quickstart",
+ "sourceUrl": "https://github.com/InferMatrixLab/InferRSI/blob/main/docs/quickstart.md",
+ "sourceNote": "Verified as the official quickstart for a workbench environment with the target engine already prepared; it does not install the target model or GPU environment. The API modification timestamp is not used as a publication date; no individual author is asserted.",
+ "verifiedAt": "2026-10-30"
+ },
+ {
+ "id": "inferrsi-qwen-image-b300-case",
+ "project": "inferrsi",
+ "type": "technical",
+ "title": {
+ "zh": "InferRSI 优化案例:Qwen-Image-2.1-Turbo 在 B300 上的验证过程",
+ "en": "InferRSI Case Study: Validating Qwen-Image-2.1-Turbo Optimizations on B300"
+ },
+ "description": {
+ "zh": "阅读固定工作负载下的优化记录、交替 A/B 实验、图像质量验证与复现步骤,理解结果的适用范围。",
+ "en": "Follow the optimization ledger, alternating A/B trials, image validation, and reproduction steps for a fixed workload."
+ },
+ "url": "https://github.com/InferMatrixLab/InferRSI/blob/main/docs/qwen-image-2.1-turbo-b300.md",
+ "platform": "GitHub Docs",
+ "publisher": "InferMatrixLab",
+ "sourceUrl": "https://github.com/InferMatrixLab/InferRSI/blob/main/docs/qwen-image-2.1-turbo-b300.md",
+ "sourceNote": "Verified as an official experiment report with a fixed workload, optimization ledger, alternating A/B trials, image validation and reproduction instructions. Its historical acceptance limits are retained; the summary does not generalize its results into universal performance claims. The API modification timestamp is not used as a publication date; no individual author is asserted.",
+ "verifiedAt": "2026-10-30"
+ }
+]
diff --git a/index.html b/index.html
index a44b035..093b024 100644
--- a/index.html
+++ b/index.html
@@ -14,11 +14,11 @@
-
+
-
+
@@ -33,7 +33,8 @@
品牌愿景开源项目演示展示
- 视频资源
+ 视频专栏
+ 文章专栏InferRSIRecipes ↗一起构建
@@ -132,9 +133,14 @@