{
  "articles": [
    {
      "id": "item-1tz12t2",
      "kind": "article",
      "title": "Helping K–12 educators build practical AI skills",
      "url": "https://openai.com/index/k-12-educators-practical-skills",
      "source": "OpenAI 新闻",
      "sourceUrl": "https://openai.com/news/",
      "publishedAt": "2026-07-08T10:00:00.000Z",
      "summary": "来自 OpenAI 新闻 的文章，主题是“Helping K–12 educators build practical AI skills”。适合边看边动手，重点看核心概念、典型场景和可复用步骤，学完可以尝试整理 3 条可马上复用的实践笔记。",
      "category": "OpenAI",
      "level": "实战",
      "minutes": 8,
      "score": 39
    },
    {
      "id": "item-1lmktjl",
      "kind": "article",
      "title": "Profiling in PyTorch (Part 3)：Attention is all you profile",
      "url": "https://huggingface.co/blog/torch-attention-profile",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-10T00:00:00.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“Profiling in PyTorch (Part 3)：Attention is all you profile”。适合补齐进阶细节，重点看训练、推理和框架底层优化，学完可以尝试复现一个最小实验并观察关键参数。",
      "category": "深度学习",
      "level": "进阶",
      "minutes": 7,
      "score": 26
    },
    {
      "id": "item-40krb5",
      "kind": "article",
      "title": "PyTorch 2.13 Release Blog",
      "url": "https://pytorch.org/blog/pytorch-2-13-release-blog/",
      "source": "PyTorch 博客",
      "sourceUrl": "https://pytorch.org/blog/",
      "publishedAt": "2026-07-08T17:42:23.000Z",
      "summary": "来自 PyTorch 博客 的文章，主题是“PyTorch 2.13 Release Blog”。适合补齐进阶细节，重点看训练、推理和框架底层优化，学完可以尝试复现一个最小实验并观察关键参数。",
      "category": "深度学习",
      "level": "进阶",
      "minutes": 8,
      "score": 20
    },
    {
      "id": "item-o3fc1p",
      "kind": "article",
      "title": "Data 面向 智能体",
      "url": "https://huggingface.co/blog/nvidia/open-data-for-agents",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-08T17:16:05.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“Data 面向 智能体”。适合补齐进阶细节，重点看智能体的工具调用、任务拆解和可靠性，学完可以尝试拆一个带工具调用的智能体流程。",
      "category": "Agent",
      "level": "进阶",
      "minutes": 7,
      "score": 20
    },
    {
      "id": "item-14ppbn8",
      "kind": "article",
      "title": "Native-speed vLLM transformers modeling backend",
      "url": "https://huggingface.co/blog/native-speed-vllm-transformers-backend",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-08T00:00:00.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“Native-speed vLLM transformers modeling backend”。适合补齐进阶细节，重点看推理部署、性能压测和上线前评估，学完可以尝试记录一次模型部署的延迟和吞吐指标。",
      "category": "部署",
      "level": "进阶",
      "minutes": 7,
      "score": 13
    },
    {
      "id": "item-1n3kdp0",
      "kind": "article",
      "title": "从 Hugging Face 到 Amazon SageMaker Studio in one click",
      "url": "https://huggingface.co/blog/amazon/one-click-to-sagemaker-studio",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-07T21:15:33.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“从 Hugging Face 到 Amazon SageMaker Studio in one click”。适合补齐进阶细节，重点看核心概念、典型场景和可复用步骤，学完可以尝试整理 3 条可马上复用的实践笔记。",
      "category": "开源模型",
      "level": "进阶",
      "minutes": 7,
      "score": 20
    },
    {
      "id": "item-gm9k8c",
      "kind": "article",
      "title": "Hugging Face Models on Foundry Managed Compute",
      "url": "https://huggingface.co/blog/microsoft/foundry-managed-compute",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-07T15:20:06.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“Hugging Face Models on Foundry Managed Compute”。适合补齐进阶细节，重点看核心概念、典型场景和可复用步骤，学完可以尝试整理 3 条可马上复用的实践笔记。",
      "category": "开源模型",
      "level": "进阶",
      "minutes": 7,
      "score": 27
    },
    {
      "id": "item-1xhsjwn",
      "kind": "article",
      "title": "Run AI workloads on any cloud, store on Hugging Face：zero-egress storage 结合 SkyPilot",
      "url": "https://huggingface.co/blog/skypilot-hf-storage",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-07T00:00:00.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“Run AI workloads on any cloud, store on Hugging Face：zero-egress storage 结合 SkyPilot”。适合补齐进阶细节，重点看检索、嵌入向量与知识库的连接方式，学完可以尝试做一个小型语义搜索或知识库问答演示项目。",
      "category": "RAG",
      "level": "进阶",
      "minutes": 7,
      "score": 27
    },
    {
      "id": "item-b7h5md",
      "kind": "article",
      "title": "Bringing PyTorch Monarch 到 AMD GPUs：Single-Controller Distributed Training on ROCm",
      "url": "https://pytorch.org/blog/bringing-pytorch-monarch-to-amd-gpus-single-controller-distributed-training-on-rocm/",
      "source": "PyTorch 博客",
      "sourceUrl": "https://pytorch.org/blog/",
      "publishedAt": "2026-07-06T23:31:47.000Z",
      "summary": "来自 PyTorch 博客 的文章，主题是“Bringing PyTorch Monarch 到 AMD GPUs：Single-Controller Distributed Training on ROCm”。适合补齐进阶细节，重点看训练、推理和框架底层优化，学完可以尝试复现一个最小实验并观察关键参数。",
      "category": "深度学习",
      "level": "进阶",
      "minutes": 8,
      "score": 27
    },
    {
      "id": "item-1v3myqt",
      "kind": "article",
      "title": "Hugging Face 与 Cerebras bring Gemma 4 到 real-time voice AI",
      "url": "https://huggingface.co/blog/cerebras-gemma4-voice-ai",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-01T00:00:00.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“Hugging Face 与 Cerebras bring Gemma 4 到 real-time voice AI”。适合补齐进阶细节，重点看核心概念、典型场景和可复用步骤，学完可以尝试整理 3 条可马上复用的实践笔记。",
      "category": "开源模型",
      "level": "进阶",
      "minutes": 7,
      "score": 26
    },
    {
      "id": "item-qp8q86",
      "kind": "article",
      "title": "ScarfBench：压测 AI 智能体 面向 Enterprise Java Framework Migration",
      "url": "https://huggingface.co/blog/ibm-research/scarfbench",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-06-30T18:32:50.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“ScarfBench：压测 AI 智能体 面向 Enterprise Java Framework Migration”。适合深入理解方法，重点看智能体的工具调用、任务拆解和可靠性，学完可以尝试拆一个带工具调用的智能体流程。",
      "category": "Agent",
      "level": "研究",
      "minutes": 7,
      "score": 26
    },
    {
      "id": "item-62x8h0",
      "kind": "article",
      "title": "LeRobot v0.6.0：Imagine, Evaluate, Improve",
      "url": "https://huggingface.co/blog/lerobot-release-v060",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-07T00:00:00.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“LeRobot v0.6.0：Imagine, Evaluate, Improve”。适合补齐进阶细节，重点看核心概念、典型场景和可复用步骤，学完可以尝试整理 3 条可马上复用的实践笔记。",
      "category": "开源模型",
      "level": "进阶",
      "minutes": 7,
      "score": 13
    },
    {
      "id": "item-14zl1mr",
      "kind": "article",
      "title": "PRX Part 4：Our Data Strategy",
      "url": "https://huggingface.co/blog/Photoroom/prx-part4-data",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-06T15:30:55.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“PRX Part 4：Our Data Strategy”。适合补齐进阶细节，重点看核心概念、典型场景和可复用步骤，学完可以尝试整理 3 条可马上复用的实践笔记。",
      "category": "开源模型",
      "level": "进阶",
      "minutes": 7,
      "score": 13
    },
    {
      "id": "item-1rg68ul",
      "kind": "article",
      "title": "🤗 Kernels：Major Updates",
      "url": "https://huggingface.co/blog/revamped-kernels",
      "source": "Hugging Face 博客",
      "sourceUrl": "https://huggingface.co/blog",
      "publishedAt": "2026-07-06T00:00:00.000Z",
      "summary": "来自 Hugging Face 博客 的文章，主题是“🤗 Kernels：Major Updates”。适合补齐进阶细节，重点看训练、推理和框架底层优化，学完可以尝试整理 3 条可马上复用的实践笔记。",
      "category": "开源模型",
      "level": "进阶",
      "minutes": 7,
      "score": 13
    },
    {
      "id": "item-l61xcr",
      "kind": "article",
      "title": "Understanding PyTorch’s Test Infrastructure",
      "url": "https://pytorch.org/blog/understanding-pytorchs-test-infrastructure/",
      "source": "PyTorch 博客",
      "sourceUrl": "https://pytorch.org/blog/",
      "publishedAt": "2026-07-03T15:00:07.000Z",
      "summary": "来自 PyTorch 博客 的文章，主题是“Understanding PyTorch’s Test Infrastructure”。适合补齐进阶细节，重点看训练、推理和框架底层优化，学完可以尝试复现一个最小实验并观察关键参数。",
      "category": "深度学习",
      "level": "进阶",
      "minutes": 8,
      "score": 20
    },
    {
      "id": "item-5klg82",
      "kind": "article",
      "title": "Building the Future of On-Device AI at the ExecuTorch Hackathon",
      "url": "https://pytorch.org/blog/building-the-future-of-on-device-ai-at-the-executorch-hackathon/",
      "source": "PyTorch 博客",
      "sourceUrl": "https://pytorch.org/blog/",
      "publishedAt": "2026-07-02T20:47:16.000Z",
      "summary": "来自 PyTorch 博客 的文章，主题是“Building the Future of On-Device AI at the ExecuTorch Hackathon”。适合深入理解方法，重点看训练、推理和框架底层优化，学完可以尝试复现一个最小实验并观察关键参数。",
      "category": "深度学习",
      "level": "研究",
      "minutes": 8,
      "score": 25
    },
    {
      "id": "item-1uht58v",
      "kind": "article",
      "title": "Shopify Joins the PyTorch Foundation as a Platinum Member",
      "url": "https://pytorch.org/blog/shopify-joins-the-pytorch-foundation-as-a-platinum-member/",
      "source": "PyTorch 博客",
      "sourceUrl": "https://pytorch.org/blog/",
      "publishedAt": "2026-07-01T15:55:59.000Z",
      "summary": "来自 PyTorch 博客 的文章，主题是“Shopify Joins the PyTorch Foundation as a Platinum Member”。适合补齐进阶细节，重点看训练、推理和框架底层优化，学完可以尝试复现一个最小实验并观察关键参数。",
      "category": "深度学习",
      "level": "进阶",
      "minutes": 8,
      "score": 26
    },
    {
      "id": "item-1j7qw9h",
      "kind": "article",
      "title": "Miles：A PyTorch-Native Stack 面向 Large-Scale 大语言模型 RL Post-Training",
      "url": "https://pytorch.org/blog/miles-a-pytorch-native-stack-for-large-scale-llm-rl-post-training/",
      "source": "PyTorch 博客",
      "sourceUrl": "https://pytorch.org/blog/",
      "publishedAt": "2026-06-30T17:48:27.000Z",
      "summary": "来自 PyTorch 博客 的文章，主题是“Miles：A PyTorch-Native Stack 面向 Large-Scale 大语言模型 RL Post-Training”。适合补齐进阶细节，重点看训练、推理和框架底层优化，学完可以尝试复现一个最小实验并观察关键参数。",
      "category": "深度学习",
      "level": "进阶",
      "minutes": 8,
      "score": 26
    }
  ],
  "videos": [
    {
      "id": "item-uvlyyk",
      "kind": "video",
      "title": "Voice 面向 AI 智能体 与 Applications",
      "url": "https://www.youtube.com/watch?v=z4gueKVs6Bo",
      "source": "DeepLearningAI",
      "sourceUrl": "https://www.youtube.com/@Deeplearningai",
      "publishedAt": "2026-06-17T15:00:38.000Z",
      "summary": "来自 DeepLearningAI 的视频，主题是“Voice 面向 AI 智能体 与 Applications”。适合边看边动手，重点看智能体的工具调用、任务拆解和可靠性，学完可以尝试拆一个带工具调用的智能体流程。",
      "category": "Agent",
      "level": "实战",
      "minutes": 11,
      "thumbnail": "https://i3.ytimg.com/vi/z4gueKVs6Bo/hqdefault.jpg",
      "score": 46
    },
    {
      "id": "item-npcmj0",
      "kind": "video",
      "title": "7-day Voice AI Build Challenge",
      "url": "https://www.youtube.com/shorts/7D3w87YqfA8",
      "source": "DeepLearningAI",
      "sourceUrl": "https://www.youtube.com/@Deeplearningai",
      "publishedAt": "2026-06-24T17:11:15.000Z",
      "summary": "来自 DeepLearningAI 的视频，主题是“7-day Voice AI Build Challenge”。适合边看边动手，重点看智能体的工具调用、任务拆解和可靠性，学完可以尝试拆一个带工具调用的智能体流程。",
      "category": "Agent",
      "level": "实战",
      "minutes": 11,
      "thumbnail": "https://i4.ytimg.com/vi/7D3w87YqfA8/hqdefault.jpg",
      "score": 33
    },
    {
      "id": "item-1dz0itg",
      "kind": "video",
      "title": "语义搜索从嵌入向量开始",
      "url": "https://www.youtube.com/shorts/S_xwTgs8zmo",
      "source": "DeepLearningAI",
      "sourceUrl": "https://www.youtube.com/@Deeplearningai",
      "publishedAt": "2026-05-22T19:12:39.000Z",
      "summary": "来自 DeepLearningAI 的视频，主题是“语义搜索从嵌入向量开始”。适合边看边动手，重点看检索、嵌入向量与知识库的连接方式，学完可以尝试做一个小型语义搜索或知识库问答演示项目。",
      "category": "RAG",
      "level": "实战",
      "minutes": 11,
      "thumbnail": "https://i4.ytimg.com/vi/S_xwTgs8zmo/hqdefault.jpg",
      "score": 35
    }
  ],
  "updatedAt": "2026-07-10T10:48:56.221Z",
  "generatedBy": "live",
  "sourceCount": 9,
  "failures": [
    "Cloudflare AI: HTTP 429"
  ]
}