{
  "schemaVersion": "1.0",
  "item": {
    "slug": "gpu-deploy",
    "name": "Gpu Deploy",
    "source": "tencent",
    "type": "skill",
    "category": "AI 智能",
    "sourceUrl": "https://clawhub.ai/wang-junjian/gpu-deploy",
    "canonicalUrl": "https://clawhub.ai/wang-junjian/gpu-deploy",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/gpu-deploy",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=gpu-deploy",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "README.md",
      "SKILL.md"
    ],
    "primaryDoc": "SKILL.md",
    "quickSetup": [
      "Download the package from Yavira.",
      "Extract the archive and review SKILL.md first.",
      "Import or place the package into your OpenClaw setup."
    ],
    "agentAssist": {
      "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
      "steps": [
        "Download the package from Yavira.",
        "Extract it into a folder your agent can access.",
        "Paste one of the prompts below and point your agent at the extracted folder."
      ],
      "prompts": [
        {
          "label": "New install",
          "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete."
        },
        {
          "label": "Upgrade existing",
          "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
        }
      ]
    },
    "sourceHealth": {
      "source": "tencent",
      "slug": "gpu-deploy",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-05-01T18:53:32.197Z",
      "expiresAt": "2026-05-08T18:53:32.197Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=gpu-deploy",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=gpu-deploy",
        "contentDisposition": "attachment; filename=\"gpu-deploy-0.1.0.zip\"",
        "redirectLocation": null,
        "bodySnippet": null,
        "slug": "gpu-deploy"
      },
      "scope": "item",
      "summary": "Item download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this item.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/gpu-deploy"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    },
    "downloadPageUrl": "https://openagent3.xyz/downloads/gpu-deploy",
    "agentPageUrl": "https://openagent3.xyz/skills/gpu-deploy/agent",
    "manifestUrl": "https://openagent3.xyz/skills/gpu-deploy/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/gpu-deploy/agent.md"
  },
  "agentAssist": {
    "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
    "steps": [
      "Download the package from Yavira.",
      "Extract it into a folder your agent can access.",
      "Paste one of the prompts below and point your agent at the extracted folder."
    ],
    "prompts": [
      {
        "label": "New install",
        "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete."
      },
      {
        "label": "Upgrade existing",
        "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run."
      }
    ]
  },
  "documentation": {
    "source": "clawhub",
    "primaryDoc": "SKILL.md",
    "sections": [
      {
        "title": "🚀 GPU 部署技能",
        "body": "在 GPU 服务器上快速部署 vLLM 模型服务。"
      },
      {
        "title": "✨ 功能特点",
        "body": "🖥️ 多服务器支持 - 配置多个 GPU 服务器，灵活选择\n🔍 自动检查 - 一键检查 GPU 状态和端口占用\n🤖 模型库 - 预置流行模型配置\n⚡ 快速部署 - 简单命令即可启动服务"
      },
      {
        "title": "1. 配置服务器",
        "body": "创建 ~/.config/gpu-deploy/servers.json：\n\n{\n  \"servers\": {\n    \"gpu1\": {\n      \"host\": \"gpu1\",\n      \"user\": \"lnsoft\",\n      \"gpu_count\": 4,\n      \"model_path\": \"/data/models/llm\"\n    },\n    \"my-gpu\": {\n      \"host\": \"192.168.1.100\",\n      \"user\": \"ubuntu\",\n      \"gpu_count\": 2,\n      \"model_path\": \"/home/ubuntu/models\"\n    }\n  },\n  \"default_server\": \"gpu1\"\n}"
      },
      {
        "title": "2. 检查服务器状态",
        "body": "# 使用默认服务器\ngpu-deploy check\n\n# 指定服务器\ngpu-deploy check --server gpu1"
      },
      {
        "title": "3. 部署模型",
        "body": "# 部署预设模型\ngpu-deploy deploy deepseek-r1-32b\n\n# 指定端口\ngpu-deploy deploy deepseek-r1-32b --port 8112"
      },
      {
        "title": "check - 检查服务器状态",
        "body": "检查 GPU 显存和端口占用情况。\n\ngpu-deploy check [--server NAME] [--port PORT]\n\n输出示例：\n\n✅ GPU 状态正常\n- 4 × Tesla T4 (15GB)\n- 显存占用: 12.6GB/卡\n- 温度: 51-55°C\n\n✅ 端口 8111 可用"
      },
      {
        "title": "deploy - 部署模型",
        "body": "启动 vLLM 模型服务。\n\ngpu-deploy deploy <MODEL_NAME> [--server NAME] [--port PORT]\n\n支持的模型：\n\ndeepseek-r1-32b - DeepSeek-R1-Distill-Qwen-32B-AWQ\nllama-3-8b - Llama 3 8B\nqwen-7b - Qwen 7B\nmistral-7b - Mistral 7B"
      },
      {
        "title": "list - 列出可用模型",
        "body": "gpu-deploy list"
      },
      {
        "title": "ps - 查看运行中的服务",
        "body": "gpu-deploy ps [--server NAME]"
      },
      {
        "title": "stop - 停止服务",
        "body": "gpu-deploy stop [--server NAME] [--port PORT]"
      },
      {
        "title": "🔧 手动使用（无脚本）",
        "body": "如果不想用封装脚本，也可以直接用原始命令："
      },
      {
        "title": "检查 GPU",
        "body": "ssh <user>@<host> nvidia-smi"
      },
      {
        "title": "检查端口",
        "body": "ssh <user>@<host> \"lsof -i :<port> 2>/dev/null || echo '端口可用'\""
      },
      {
        "title": "部署模型（DeepSeek R1 32B）",
        "body": "ssh <user>@<host> \"tmux new-session -d -s vllm '\nsource /data/miniconda3/etc/profile.d/conda.sh && \\\nconda activate vllm && \\\ncd /data/models/llm && \\\nvllm serve /data/models/llm/deepseek/DeepSeek-R1-Distill-Qwen-32B-AWQ/ \\\n  --tensor-parallel-size 4 \\\n  --max-model-len 102400 \\\n  --dtype half \\\n  --port 8111 \\\n  --served-model-name gpt-4o-mini\n'\""
      },
      {
        "title": "📦 添加自定义模型",
        "body": "在 ~/.config/gpu-deploy/models.json 中添加：\n\n{\n  \"my-model\": {\n    \"name\": \"My Awesome Model\",\n    \"path\": \"/path/to/model\",\n    \"tensor_parallel_size\": 2,\n    \"max_model_len\": 8192,\n    \"dtype\": \"half\",\n    \"port\": 8111,\n    \"served_model_name\": \"my-model\"\n  }\n}"
      },
      {
        "title": "⚠️ 注意事项",
        "body": "部署前检查 - 总是先运行 check 确认资源可用\n后台运行 - 建议使用 tmux/screen 保持服务运行\n端口管理 - 不同模型使用不同端口\n显存估算 - 7B 模型约需 8-10GB，32B 约需 10-14GB/卡"
      },
      {
        "title": "🔗 相关链接",
        "body": "vLLM 文档: https://docs.vllm.ai\n模型下载: https://huggingface.co/models\n问题反馈: https://github.com/your-username/gpu-deploy-skill\n\n由 OpenClaw 社区贡献 🦞"
      }
    ],
    "body": "🚀 GPU 部署技能\n\n在 GPU 服务器上快速部署 vLLM 模型服务。\n\n✨ 功能特点\n🖥️ 多服务器支持 - 配置多个 GPU 服务器，灵活选择\n🔍 自动检查 - 一键检查 GPU 状态和端口占用\n🤖 模型库 - 预置流行模型配置\n⚡ 快速部署 - 简单命令即可启动服务\n📋 快速开始\n1. 配置服务器\n\n创建 ~/.config/gpu-deploy/servers.json：\n\n{\n  \"servers\": {\n    \"gpu1\": {\n      \"host\": \"gpu1\",\n      \"user\": \"lnsoft\",\n      \"gpu_count\": 4,\n      \"model_path\": \"/data/models/llm\"\n    },\n    \"my-gpu\": {\n      \"host\": \"192.168.1.100\",\n      \"user\": \"ubuntu\",\n      \"gpu_count\": 2,\n      \"model_path\": \"/home/ubuntu/models\"\n    }\n  },\n  \"default_server\": \"gpu1\"\n}\n\n2. 检查服务器状态\n# 使用默认服务器\ngpu-deploy check\n\n# 指定服务器\ngpu-deploy check --server gpu1\n\n3. 部署模型\n# 部署预设模型\ngpu-deploy deploy deepseek-r1-32b\n\n# 指定端口\ngpu-deploy deploy deepseek-r1-32b --port 8112\n\n🎛️ 可用命令\ncheck - 检查服务器状态\n\n检查 GPU 显存和端口占用情况。\n\ngpu-deploy check [--server NAME] [--port PORT]\n\n\n输出示例：\n\n✅ GPU 状态正常\n- 4 × Tesla T4 (15GB)\n- 显存占用: 12.6GB/卡\n- 温度: 51-55°C\n\n✅ 端口 8111 可用\n\ndeploy - 部署模型\n\n启动 vLLM 模型服务。\n\ngpu-deploy deploy <MODEL_NAME> [--server NAME] [--port PORT]\n\n\n支持的模型：\n\ndeepseek-r1-32b - DeepSeek-R1-Distill-Qwen-32B-AWQ\nllama-3-8b - Llama 3 8B\nqwen-7b - Qwen 7B\nmistral-7b - Mistral 7B\nlist - 列出可用模型\ngpu-deploy list\n\nps - 查看运行中的服务\ngpu-deploy ps [--server NAME]\n\nstop - 停止服务\ngpu-deploy stop [--server NAME] [--port PORT]\n\n🔧 手动使用（无脚本）\n\n如果不想用封装脚本，也可以直接用原始命令：\n\n检查 GPU\nssh <user>@<host> nvidia-smi\n\n检查端口\nssh <user>@<host> \"lsof -i :<port> 2>/dev/null || echo '端口可用'\"\n\n部署模型（DeepSeek R1 32B）\nssh <user>@<host> \"tmux new-session -d -s vllm '\nsource /data/miniconda3/etc/profile.d/conda.sh && \\\nconda activate vllm && \\\ncd /data/models/llm && \\\nvllm serve /data/models/llm/deepseek/DeepSeek-R1-Distill-Qwen-32B-AWQ/ \\\n  --tensor-parallel-size 4 \\\n  --max-model-len 102400 \\\n  --dtype half \\\n  --port 8111 \\\n  --served-model-name gpt-4o-mini\n'\"\n\n📦 添加自定义模型\n\n在 ~/.config/gpu-deploy/models.json 中添加：\n\n{\n  \"my-model\": {\n    \"name\": \"My Awesome Model\",\n    \"path\": \"/path/to/model\",\n    \"tensor_parallel_size\": 2,\n    \"max_model_len\": 8192,\n    \"dtype\": \"half\",\n    \"port\": 8111,\n    \"served_model_name\": \"my-model\"\n  }\n}\n\n⚠️ 注意事项\n部署前检查 - 总是先运行 check 确认资源可用\n后台运行 - 建议使用 tmux/screen 保持服务运行\n端口管理 - 不同模型使用不同端口\n显存估算 - 7B 模型约需 8-10GB，32B 约需 10-14GB/卡\n🔗 相关链接\nvLLM 文档: https://docs.vllm.ai\n模型下载: https://huggingface.co/models\n问题反馈: https://github.com/your-username/gpu-deploy-skill\n\n由 OpenClaw 社区贡献 🦞"
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/wang-junjian/gpu-deploy",
    "publisherUrl": "https://clawhub.ai/wang-junjian/gpu-deploy",
    "owner": "wang-junjian",
    "version": "0.1.0",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/gpu-deploy",
    "downloadUrl": "https://openagent3.xyz/downloads/gpu-deploy",
    "agentUrl": "https://openagent3.xyz/skills/gpu-deploy/agent",
    "manifestUrl": "https://openagent3.xyz/skills/gpu-deploy/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/gpu-deploy/agent.md"
  }
}