Logging MLflow Dictionary Artifacts in Azure Machine Learning

DP-100 By islever Editorial Last updated Aug 21, 2026

Short answer

C mlflow.log_metrics(my_dict)

Question

You manage an Azure Machine Learning workspace. You experiment with an MLflow model that trains interactively by using a notebook in the workspace. You need to log dictionary type artifacts of the experiments in Azure Machine Learning by using MLflow. Which syntax should you use?

mlflow.log_input(my_dict)
mlflow.log_metric("my_metric", my_dict)
mlflow.log_metrics(my_dict) Correct Answer
mlflow.log_text("my_metric", my_dict)

Key Takeaways

Tests knowledge of MLflow tracking functions; the trap is confusing metric logging with parameter or artifact logging, though the question's use of 'artifacts' is technically imprecise terminology for metrics.

This question tests the correct MLflow Python API syntax for logging multiple key-value pairs as metrics. The community consensus confirms that mlflow.log_metrics() is the standard method for handling dictionary inputs in ML experiments.

Common mistake: Option B (log_metric) is incorrect because it expects a single value, not a dictionary. Option A (log_input) is for input data tracking, and D (log_text) is for raw text strings.

Expert Analysis

Why the Answer Is Correct

The mlflow.log_metrics() function accepts a dictionary of key-value pairs where keys are metric names and values are numeric scores. This allows efficient batch logging of multiple evaluation results (e.g., accuracy, loss) in a single call, which matches the requirement to log a dictionary.

Why the Other Options Are Wrong

mlflow.log_metric takes a single string key and float value, making it unsuitable for dictionaries. mlflow.log_input is used to track dataset versions, not model performance metrics. mlflow.log_text logs arbitrary text content, not structured numerical metrics.

Community Comment Notes

Comment [1] provides official documentation examples showing mlflow.log_metrics(metrics) with a dictionary. Comment [3] correctly notes that if these were hyperparameters, log_params would be appropriate, but since they are typically experiment outputs, log_metrics is the best fit among the choices.

Official Reference

https://mlflow.org/docs/latest/python_api/mlflow.html#mlflow.log_metrics

This link opens the official vendor documentation in a new tab.

Exam Strategy

Distinguish between MLflow functions by their data type: log_metric/log_metrics for numbers, log_param/log_params for strings/configs, and log_artifact for files. Always check if the input is a single value or a collection to choose between singular and plural method names.

Community Votes

100%

100% of anonymous learners picked answer C. Votes are pick records left by other test-takers — they are not the verified answer.

Discussion 4 contributions

Compiled from public community threads and reader submissions, lightly edited for clarity. Handles and vote tallies are shown as recorded, and comments reflect the views of individual contributors.

From public community threads 4

avinyc 👍 2 Answered C

Ref - https://mlflow.org/docs/latest/python_api/mlflow.html#mlflow.log_input import mlflow metrics = {"mse": 2500.00, "rmse": 50.00} # Log a batch of metrics with mlflow.start_run(): mlflow.log_metrics(metrics) # Log a batch of metrics in async fashion. with mlflow.start_run(): mlflow.log_metrics(metrics, synchronous=False)
jl420 👍 2

Answer is C - mlflow.log_metrics(my_dict) Explanation: In MLflow, when you want to log dictionary-type artifacts that contain multiple metrics, you should use the mlflow.log_metrics() method. This method allows you to log multiple key-value pairs (i.e., a dictionary) as metrics.
AzureGeek79 👍 2

the correct answer is log_params which is not included in the available options
jefimija 👍 1

log_artifact maybe?

Reader comments & corrections

No reader corrections yet — spotted an error or have a note? Share it below.

← Back to DP-100 Study Guide