Query & compare experiments and runs with MLflow

Experiments and jobs (or runs) in Azure Machine Learning can be queried using MLflow. You don't need to install any specific SDK to manage what happens inside of a training job, creating a more seamless transition between local runs and the cloud by removing cloud-specific dependencies. In this article, you'll learn how to query and compare experiments and runs in your workspace using Azure Machine Learning and MLflow SDK in Python.

MLflow allows you to:

• Create, query, delete, and search for experiments in a workspace.
• Query, delete, and search for runs in a workspace.
• Track and retrieve metrics, parameters, artifacts, and models from runs.

For a detailed comparison between open-source MLflow and MLflow when connected to Azure Machine Learning, see Support matrix for querying runs and experiments in Azure Machine Learning.

You can also query and search experiments and runs by using the MLflow REST API. See Using MLflow REST with Azure Machine Learning for an example about how to consume it.

Prerequisites

• Install the MLflow SDK mlflow package and the Azure Machine Learning azureml-mlflow plugin for MLflow as follows:

  pip install mlflow azureml-mlflow
  

  Tip

  You can use the mlflow-skinny package, which is a lightweight MLflow package without SQL storage, server, UI, or data science dependencies. This package is recommended for users who primarily need the MLflow tracking and logging capabilities without importing the full suite of features, including deployments.

• Create an Azure Machine Learning workspace. To create a workspace, see Create resources you need to get started. Review the access permissions you need to perform your MLflow operations in your workspace.

• To do remote tracking, or track experiments running outside Azure Machine Learning, configure MLflow to point to the tracking URI of your Azure Machine Learning workspace. For more information on how to connect MLflow to your workspace, see Configure MLflow for Azure Machine Learning.

Query and search experiments

Use MLflow to search for experiments inside of your workspace. See the following examples:

• Get all active experiments:

  mlflow.search_experiments()
  

  Note

  In legacy versions of MLflow (<2.0), use method mlflow.list_experiments() instead.

• Get all the experiments, including archived:

  from mlflow.entities import ViewType
  
  mlflow.search_experiments(view_type=ViewType.ALL)
  

• Get a specific experiment by name:

  mlflow.get_experiment_by_name(experiment_name)
  

• Get a specific experiment by ID:

  mlflow.get_experiment('1234-5678-90AB-CDEFG')
  

Search experiments

The search_experiments() method, available since Mlflow 2.0, lets you search for experiments that match criteria using filter_string.

• Retrieve multiple experiments based on their IDs:

  mlflow.search_experiments(filter_string="experiment_id IN ("
      "'CDEFG-1234-5678-90AB', '1234-5678-90AB-CDEFG', '5678-1234-90AB-CDEFG')"
  )
  

• Retrieve all experiments created after a given time:

  import datetime
  
  dt = datetime.datetime(2022, 6, 20, 5, 32, 48)
  mlflow.search_experiments(filter_string=f"creation_time > {int(dt.timestamp())}")
  

• Retrieve all experiments with a given tag:

  mlflow.search_experiments(filter_string=f"tags.framework = 'torch'")
  

Query and search runs

MLflow lets you search for runs inside any experiment, including multiple experiments at the same time. The method mlflow.search_runs() accepts the argument experiment_ids and experiment_name to indicate which experiments you want to search. You can also indicate search_all_experiments=True if you want to search across all the experiments in the workspace:

• By experiment name:

  mlflow.search_runs(experiment_names=[ "my_experiment" ])
  

• By experiment ID:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ])
  

• Search across all experiments in the workspace:

  mlflow.search_runs(filter_string="params.num_boost_round='100'", search_all_experiments=True)
  

Notice that experiment_ids supports providing an array of experiments, so you can search runs across multiple experiments, if necessary. This might be useful in case you want to compare runs of the same model when it's being logged in different experiments (for example, by different people or different project iterations).

By default, MLflow returns the data in Pandas Dataframe format, which makes it handy when doing further processing our analysis of the runs. Returned data includes columns with:

• Basic information about the run.
• Parameters with column's name params.<parameter-name>.
• Metrics (last logged value of each) with column's name metrics.<metric-name>.

All metrics and parameters are also returned when querying runs. However, for metrics that contain multiple values (for instance, a loss curve, or a PR curve), only the last value of the metric is returned. If you want to retrieve all the values of a given metric, uses mlflow.get_metric_history method. See Getting params and metrics from a run for an example.

Order runs

By default, experiments are in descending order by start_time, which is the time the experiment was queued in Azure Machine Learning. However, you can change this default by using the parameter order_by.

• Order runs by attributes, like start_time:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ],
                     order_by=["attributes.start_time DESC"])
  

• Order runs and limit results. The following example returns the last single run in the experiment:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     max_results=1, order_by=["attributes.start_time DESC"])
  

• Order runs by the attribute duration:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     order_by=["attributes.duration DESC"])
  

  Tip

  attributes.duration isn't present in MLflow OSS, but provided in Azure Machine Learning for convenience.

• Order runs by metric's values:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ]).sort_values("metrics.accuracy", ascending=False)
  

  Warning

  Using order_by with expressions containing metrics.*, params.*, or tags.* in the parameter order_by isn't currently supported. Instead, use the sort_values method from Pandas as shown in the example.

Filter runs

You can also look for a run with a specific combination in the hyperparameters using the parameter filter_string. Use params to access run's parameters, metrics to access metrics logged in the run, and attributes to access run information details. MLflow supports expressions joined by the AND keyword (the syntax doesn't support OR):

• Search runs based on a parameter's value:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string="params.num_boost_round='100'")
  

  Warning

  Only operators =, like, and != are supported for filtering parameters.

• Search runs based on a metric's value:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string="metrics.auc>0.8")
  

• Search runs with a given tag:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string="tags.framework='torch'")
  

• Search runs created by a given user:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string="attributes.user_id = 'John Smith'")
  

• Search runs that failed. See Filter runs by status for possible values:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string="attributes.status = 'Failed'")
  

• Search runs created after a given time:

  import datetime
  
  dt = datetime.datetime(2022, 6, 20, 5, 32, 48)
  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string=f"attributes.creation_time > '{int(dt.timestamp())}'")
  

  Tip

  For the key attributes, values should always be strings and hence encoded between quotes.

• Search runs that take longer than one hour:

  duration = 360 * 1000 # duration is in milliseconds
  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string=f"attributes.duration > '{duration}'")
  

  Tip

  attributes.duration isn't present in MLflow OSS, but provided in Azure Machine Learning for convenience.

• Search runs that have the ID in a given set:

  mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                     filter_string="attributes.run_id IN ('1234-5678-90AB-CDEFG', '5678-1234-90AB-CDEFG')")
  

Filter runs by status

When you filter runs by status, MLflow uses a different convention to name the different possible status of a run compared to Azure Machine Learning. The following table shows the possible values:

Example:

mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                   filter_string="attributes.status = 'Failed'")

Get metrics, parameters, artifacts, and models

The method search_runs returns a Pandas Dataframe that contains a limited amount of information by default. You can get Python objects if needed, which might be useful to get details about them. Use the output_format parameter to control how output is returned:

runs = mlflow.search_runs(
    experiment_ids=[ "1234-5678-90AB-CDEFG" ],
    filter_string="params.num_boost_round='100'",
    output_format="list",
)

Details can then be accessed from the info member. The following sample shows how to get the run_id:

last_run = runs[-1]
print("Last run ID:", last_run.info.run_id)

Get params and metrics from a run

When runs are returned using output_format="list", you can easily access parameters using the key data:

last_run.data.params

In the same way, you can query metrics:

last_run.data.metrics

For metrics that contain multiple values (for instance, a loss curve, or a PR curve), only the last logged value of the metric is returned. If you want to retrieve all the values of a given metric, uses mlflow.get_metric_history method. This method requires you to use the MlflowClient:

client = mlflow.tracking.MlflowClient()
client.get_metric_history("1234-5678-90AB-CDEFG", "log_loss")

Get artifacts from a run

MLflow can query any artifact logged by a run. Artifacts can't be accessed using the run object itself, and the MLflow client should be used instead:

client = mlflow.tracking.MlflowClient()
client.list_artifacts("1234-5678-90AB-CDEFG")

The preceding method lists all the artifacts logged in the run, but they remain stored in the artifacts store (Azure Machine Learning storage). To download any of them, use the method download_artifact:

file_path = mlflow.artifacts.download_artifacts(
    run_id="1234-5678-90AB-CDEFG", artifact_path="feature_importance_weight.png"
)

Get models from a run

Models can also be logged in the run and then retrieved directly from it. To retrieve a model, you need to know the path to the artifact where it's stored. The method list_artifacts can be used to find artifacts that represent a model since MLflow models are always folders. You can download a model by specifying the path where the model is stored, using the download_artifact method:

artifact_path="classifier"
model_local_path = mlflow.artifacts.download_artifacts(
  run_id="1234-5678-90AB-CDEFG", artifact_path=artifact_path
)

You can then load the model back from the downloaded artifacts using the typical function load_model in the flavor-specific namespace. The following example uses xgboost:

model = mlflow.xgboost.load_model(model_local_path)

MLflow also allows you to perform both operations at once, and to download and load the model in a single instruction. MLflow downloads the model to a temporary folder and loads it from there. The method load_model uses an URI format to indicate from where the model has to be retrieved. In the case of loading a model from a run, the URI structure is as follows:

model = mlflow.xgboost.load_model(f"runs:/{last_run.info.run_id}/{artifact_path}")

Get child (nested) runs

MLflow supports the concept of child (nested) runs. These runs are useful when you need to spin off training routines that must be tracked independently from the main training process. Hyper-parameter tuning optimization processes or Azure Machine Learning pipelines are typical examples of jobs that generate multiple child runs. You can query all the child runs of a specific run using the property tag mlflow.parentRunId, which contains the run ID of the parent run.

hyperopt_run = mlflow.last_active_run()
child_runs = mlflow.search_runs(
    filter_string=f"tags.mlflow.parentRunId='{hyperopt_run.info.run_id}'"
)

Compare jobs and models in Azure Machine Learning studio (preview)

To compare and evaluate the quality of your jobs and models in Azure Machine Learning studio, use the preview panel to enable the feature. Once enabled, you can compare the parameters, metrics, and tags between the jobs and/or models you selected.

The MLflow with Azure Machine Learning notebooks demonstrate and expand upon concepts presented in this article.

• Train and track a classifier with MLflow: Demonstrates how to track experiments using MLflow, log models, and combine multiple flavors into pipelines.
• Manage experiments and runs with MLflow: Demonstrates how to query experiments, runs, metrics, parameters, and artifacts from Azure Machine Learning using MLflow.

Support matrix for querying runs and experiments

The MLflow SDK exposes several methods to retrieve runs, including options to control what is returned and how. Use the following table to learn about which of those methods are currently supported in MLflow when connected to Azure Machine Learning:

Related content

• Manage your models with MLflow
• Deploy models with MLflow