kubeflow-pipeline-executorSkill kubeflow-pipeline-executor

---

名称: kubeflow-pipeline-executor 描述: 用于机器学习工作流编排、组件管理和Kubernetes原生机器学习的Kubeflow Pipelines技能。 允许使用的工具:

• 读取
• 写入
• Bash
• Glob
• Grep

---

kubeflow-pipeline-executor

概述

用于机器学习工作流编排、组件管理和Kubernetes原生机器学习操作的Kubeflow Pipelines技能。

功能

• 管道定义和编译
• 组件创建和复用
• 管道版本管理
• 工件跟踪和溯源
• Kubernetes资源管理
• 管道调度和触发
• 组件输出缓存
• 管道运行可视化

目标流程

• 模型训练管道
• 分布式训练编排
• 模型部署管道
• 机器学习模型重训练管道

工具和库

• Kubeflow Pipelines
• KFP SDK (v2)
• Kubernetes
• Argo Workflows

输入模式

{
  "type": "object",
  "required": ["action"],
  "properties": {
    "action": {
      "type": "string",
      "enum": ["compile", "run", "schedule", "list", "get-run", "delete"],
      "description": "要执行的KFP操作"
    },
    "pipelinePath": {
      "type": "string",
      "description": "管道定义文件路径"
    },
    "pipelineConfig": {
      "type": "object",
      "properties": {
        "name": { "type": "string" },
        "description": { "type": "string" },
        "parameters": { "type": "object" }
      }
    },
    "runConfig": {
      "type": "object",
      "properties": {
        "experimentName": { "type": "string" },
        "runName": { "type": "string" },
        "arguments": { "type": "object" }
      }
    },
    "scheduleConfig": {
      "type": "object",
      "properties": {
        "cron": { "type": "string" },
        "maxConcurrency": { "type": "integer" },
        "enabled": { "type": "boolean" }
      }
    }
  }
}

输出模式

{
  "type": "object",
  "required": ["status", "action"],
  "properties": {
    "status": {
      "type": "string",
      "enum": ["success", "error", "running"]
    },
    "action": {
      "type": "string"
    },
    "pipelineId": {
      "type": "string"
    },
    "runId": {
      "type": "string"
    },
    "runStatus": {
      "type": "string",
      "enum": ["pending", "running", "succeeded", "failed", "skipped"]
    },
    "artifacts": {
      "type": "array",
      "items": {
        "type": "object",
        "properties": {
          "name": { "type": "string" },
          "uri": { "type": "string" },
          "type": { "type": "string" }
        }
      }
    },
    "dashboardUrl": {
      "type": "string"
    }
  }
}

使用示例

{
  kind: 'skill',
  title: '运行机器学习训练管道',
  skill: {
    name: 'kubeflow-pipeline-executor',
    context: {
      action: 'run',
      pipelinePath: 'pipelines/training_pipeline.py',
      runConfig: {
        experimentName: 'model-training',
        runName: 'training-run-v1',
        arguments: {
          dataPath: 'gs://bucket/data',
          modelPath: 'gs://bucket/models',
          epochs: 100
        }
      }
    }
  }
}