How to Use YAML Configuration¶

This guide shows you how to save, load, and validate estimator configurations as YAML files using EstimatorConfig. Use this when you want to manage estimator parameters as declarative config files rather than Python code.

Interactive notebook available

Try this guide as an interactive notebook: How to Configure Estimators with YAML

Prerequisites¶

Install the config extra:

pip install sklearn-wrap[config]

Build an Estimator from a Config¶

Create an EstimatorConfig with a dotted import path and constructor parameters, then call .build():

from sklearn_wrap.config import EstimatorConfig

config = EstimatorConfig(
    estimator_class="sklearn.linear_model.Ridge",
    params={"alpha": 2.0, "fit_intercept": True},
)

estimator = config.build()

Capture a Config from an Existing Estimator¶

Use EstimatorConfig.from_estimator() to snapshot any sklearn-compatible estimator:

from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler
from sklearn.linear_model import Ridge

pipe = Pipeline([("scaler", StandardScaler()), ("ridge", Ridge(alpha=0.5))])
config = EstimatorConfig.from_estimator(pipe)

Save and Load YAML¶

# Save
config.to_yaml("pipeline.yaml")

# Load
loaded = EstimatorConfig.from_yaml("pipeline.yaml")
estimator = loaded.build()