Saving a fitted model in Python
joblib writes the whole fitted pipeline, preprocessing included.
import joblib
from sklearn.datasets import load_iris
from sklearn.linear_model import LogisticRegression
from sklearn.pipeline import make_pipeline
from sklearn.preprocessing import StandardScaler
X, y = load_iris(return_X_y=True)
model = make_pipeline(StandardScaler(), LogisticRegression(max_iter=500))
model.fit(X, y)
joblib.dump(model, "iris.joblib")
restored = joblib.load("iris.joblib")
print(restored.predict(X[:3]))
print(f"{restored.score(X, y):.3f}")How it works
- Dump the pipeline, not just the final estimator.
- A loaded model predicts without refitting.
- Pin your library versions: a pickle is not a stable format.
Keywords and builtins used here
print
The run, in numbers
- Lines
- 15
- Characters to type
- 463
- Tokens
- 110
- Three-star pace
- 115 tpm
At the three-star pace of 115 tokens a minute, this run takes about 57 seconds.
Step 3 of 3 in Beyond the basics, step 24 of 25 in Machine learning with scikit-learn.