Automated testing for ML code in MLOps - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the testing library used for ML code.

MLOps

import [1]

Drag options to blanks, or click blank then click option'

Apytest

Btensorflow

Cnumpy

Dpandas

Attempts:

3 left

💡 Hint

Common Mistakes

Importing ML libraries instead of testing libraries.

The pytest library is commonly used for automated testing in Python, including ML code.

2fill in blank

medium

Complete the code to define a simple test function for ML model accuracy.

MLOps

def test_model_accuracy():
    accuracy = model.evaluate(X_test, y_test)[1]
    assert accuracy [1] 0.8

Drag options to blanks, or click blank then click option'

A==

B>=

C<=

D!=

Attempts:

3 left

💡 Hint

Common Mistakes

Using equality instead of comparison.

Using less than operator.

The test checks if the model accuracy is at least 0.8, so the operator should be >=.

3fill in blank

hard

Fix the error in the test assertion to correctly check model loss is below threshold.

MLOps

def test_model_loss():
    loss = model.evaluate(X_val, y_val)[0]
    assert loss [1] 0.2

Drag options to blanks, or click blank then click option'

A>=

B>

C==

D<

Attempts:

3 left

💡 Hint

Common Mistakes

Using greater than operator.

Using equality operator.

The test should assert that loss is less than 0.2, so the operator is <.

4fill in blank

hard

Fill both blanks to create a test that checks if predictions shape matches labels shape.

MLOps

def test_prediction_shape():
    preds = model.predict(X_sample)
    assert preds.[1] == y_sample.[2]

Drag options to blanks, or click blank then click option'

Ashape

Bsize

Clen

Dndim

Attempts:

3 left

💡 Hint

Common Mistakes

Using size or length which may not match dimensions.

Both predictions and labels have a shape attribute that shows their dimensions, so we compare those.

5fill in blank

hard

Fill all three blanks to write a test that filters valid predictions above threshold.

MLOps

valid_preds = [p for p in preds if p [1] [2]]
assert all(v [3] threshold for v in valid_preds)

Drag options to blanks, or click blank then click option'

A>

Bthreshold

C>=

D<

Attempts:

3 left

💡 Hint

Common Mistakes

Using wrong comparison operators.

Confusing threshold with a number.

The list comprehension filters predictions greater than the threshold, and the assertion checks all are greater or equal to it.