用 MLflow 搜索 API 构建 仪表板 Build dashboards with the MLflow Search API

你可以使用MLflow search_runs在 MLflow 上拉取聚合指标,并在仪表板中显示它们。You can pull aggregate metrics on your MLflow runs using the mlflow.search_runs API and display them in a dashboard. 定期此类审阅指标可以深入了解你的进度和工作效率。Regularly such reviewing metrics can provide insight into your progress and productivity. 例如,你可以在多个运行和/或试验中跟踪一段时间内的目标指标(如收入或准确性)的改进。For example, you can track improvement of a goal metric like revenue or accuracy over time, across many runs and/or experiments.

此笔记本演示了如何使用search_runs mlflow API 构建以下自定义仪表板:This notebook demonstrates how to build the following custom dashboard using the mlflow.search_runs API:

搜索 API 仪表板Search API Dashboard

可以在自己的试验或自动生成的模拟试验数据上运行笔记本。You can either run the notebook on your own experiments or against autogenerated mock experiment data.

比较 MLflow 运行笔记本的仪表板Dashboard comparing MLflow runs notebook

获取笔记本Get notebook