import joblib, numpy as np
from azure.ml.model import Input, Output  # If applicable

def init():
    global model
    model = joblib.load("model.joblib")

def run(data: dict):
    arr = np.array(data["data"])
    return {"predictions": model.predict(arr).tolist()}

❯ cat train.py
import os
import joblib
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier

# Load and split dataset
iris = load_iris()
X_train, X_test, y_train, y_test = train_test_split(
    iris.data, iris.target, test_size=0.2, random_state=42
)

# Train a simple classifier
clf = RandomForestClassifier(n_estimators=100, random_state=42)
clf.fit(X_train, y_train)

# Evaluate model
accuracy = clf.score(X_test, y_test)
print(f"✅ Accuracy: {accuracy:.2f}")

# Save model to ./outputs directory (Azure ML expects this as a convention)
os.makedirs("outputs", exist_ok=True)
joblib.dump(clf, "outputs/model.joblib")
print("✅ Model saved to ./outputs/model.joblib")