Implementation:Mlflow Mlflow Log Param

Knowledge Sources	MLflow MLflow API
Domains	ML_Ops, Experiment_Tracking
Last Updated	2026-02-13 20:00 GMT

Overview

Concrete tool for recording experiment hyperparameters and configuration values within an MLflow run, provided by the MLflow library.

Description

mlflow.log_param and mlflow.log_params are fluent API functions for recording key-value parameter pairs under the currently active MLflow run. log_param logs a single parameter, while log_params logs a dictionary of parameters in a single batch operation. Both convert values to strings before storage. If no run is active when either function is called, a new run is automatically created.

Usage

Use log_param for individual hyperparameters that are determined dynamically during script execution. Use log_params when logging a configuration dictionary or argparse namespace all at once. Call these functions within an active run context (inside a with mlflow.start_run() block or after calling mlflow.start_run()). Set synchronous=False in high-throughput scenarios where logging latency is a concern.

Code Reference

Source Location

Repository: mlflow
File: mlflow/tracking/fluent.py
Lines (log_param): L827-861
Lines (log_params): L1255-1298

Signature

def log_param(
    key: str,
    value: Any,
    synchronous: bool | None = None,
) -> Any: ...

def log_params(
    params: dict[str, Any],
    synchronous: bool | None = None,
    run_id: str | None = None,
) -> RunOperations | None: ...

Import

import mlflow

I/O Contract

Inputs

Name	Type	Required	Description
key	str	Yes (log_param)	Parameter name. May contain alphanumerics, underscores, dashes, periods, spaces, and slashes. Maximum length 250 characters in all backend stores.
value	Any	Yes (log_param)	Parameter value. Will be converted to a string if not already. Maximum length 6000 characters in all built-in backend stores.
params	dict[str, Any]	Yes (log_params)	Dictionary mapping parameter names to values. All values are string-ified before storage.
synchronous	bool or None	No	If True, blocks until the parameter is logged. If False, logs asynchronously and returns a future. If None, reads from `MLFLOW_ENABLE_ASYNC_LOGGING` (defaults to False).
run_id	str or None	No (log_params only)	If specified, logs parameters to the given run instead of the active run.

Outputs

Name	Type	Description
(log_param return)	Any	When `synchronous=True`, returns the parameter value as passed. When `synchronous=False`, returns a `RunOperations` instance representing the pending logging operation.
(log_params return)	RunOperations or None	When `synchronous=True`, returns None. When `synchronous=False`, returns a `RunOperations` future for the batch logging operation.

Usage Examples

Basic Usage

import mlflow

with mlflow.start_run():
    # Log individual parameters
    mlflow.log_param("learning_rate", 0.01)
    mlflow.log_param("optimizer", "adam")
    mlflow.log_param("epochs", 100)

Batch Parameter Logging

import mlflow

params = {
    "learning_rate": 0.01,
    "n_estimators": 100,
    "max_depth": 5,
    "min_samples_split": 10,
    "random_state": 42,
}

with mlflow.start_run():
    mlflow.log_params(params)

Asynchronous Logging

import mlflow

with mlflow.start_run():
    # Log asynchronously to avoid blocking the training loop
    future = mlflow.log_param("batch_size", 64, synchronous=False)

Related Pages

Implements Principle

Principle:Mlflow_Mlflow_Parameter_Logging

Requires Environment

Environment:Mlflow_Mlflow_Python_Runtime_Environment

Uses Heuristic

Page Connections

Double-click a node to navigate. Hold to expand connections.

Principle

Implementation

Heuristic

Environment