The skops library simplifies saving and loading XGBoost models.
This means you can train your model in Python and potentially deploy it in another language like Java or C++.
First, we must install the skops library using our preferred package manager, such as pip:
pip install skops
Next, we can save our XGBoost model using skops.
from sklearn.datasets import make_classification
from xgboost import XGBClassifier
from sklearn.metrics import accuracy_score
import skops.io as sio
# Generate a random classification dataset
X, y = make_classification(n_samples=1000, n_classes=2, random_state=42)
# Train an XGBoost model
model = XGBClassifier(n_estimators=100, learning_rate=0.1, random_state=42)
model.fit(X, y)
# Save the model in skops format
sio.dump(model, 'xgb_model.skops')
# Load model from file
loaded_model = sio.load('xgb_model.skops', trusted=True)
# Use the loaded model to make predictions
predictions = loaded_model.predict(X)
# Verify the loaded model performs the same as the original
print(f"Accuracy: {accuracy_score(y, predictions):.4f}")
Here’s a step-by-step breakdown:
- We train an XGBClassifier on a randomly generated binary classification dataset.
- We save the trained model in skops format using skops’
dump()function. - We load the saved model using skops’
load()function. - We use the loaded model to make predictions on the original dataset.
- We print the accuracy score to verify that the loaded model performs identically to the original model.