Evaluation output

⚠️ {{ item }}

{% for link in links %} {{ link.text }} {% endfor %}

Python APIs:

mlflow.evaluate() returns a result object that is an instance of EvaluationResult.
result.metrics: A dictionary of aggregated evaluation results.
result.tables['eval_results']: A DataFrame of per-request evaluation results.

For more details, see the mlflow.evaluate API reference or the MLflow LLM Evaluate Quickstart.