mlflow.load_mlflow_metrics

mlflow.load_mlflow_metrics(runs, *, keys=None, x='step', tracking_uri=None)

Read the metrics MLflow runs logged.

Parameters

Name	Type	Description	Default
runs	str, mlflow.entities.Run, pandas.DataFrame, or iterable of these	The runs: a run id, a run `MlflowClient.get_run()` or `search_runs()` returned, the `DataFrame` `mlflow.search_runs()` returns, or an iterable of ids and runs.	required
keys	iterable of str	Only these metrics, in this order. By default every metric a run logged, sorted, leaving out MLflow’s `system/` metrics. A metric that no run logged is warned about.	`None`
x	('step', 'relative_time')	What places a value along the x axis: the step it was logged at, or the seconds from the start of its run, as the MLflow UI offers.	`"step"`
tracking_uri	str	The tracking server to read from. By default MLflow’s own, as `mlflow.set_tracking_uri` or `MLFLOW_TRACKING_URI` set it.	`None`

Name	Type	Description
	dict of str to dict of str to MetricSeries	Metric by metric, and within a metric run by run, in the order the runs were given. A run that logged nothing under a metric is absent.

Name	Type	Description
	ImportError	If `mlflow` is not installed.
	ValueError	If `x` is neither `"step"` nor `"relative_time"`.