diff --git a/docs/source/tutorials/digit_processing.md b/docs/source/tutorials/digit_processing.md
index 2628293..a14c0be 100644
--- a/docs/source/tutorials/digit_processing.md
+++ b/docs/source/tutorials/digit_processing.md
@@ -173,14 +173,13 @@ second order (class covariances) are used to classify samples.
 
 ```{code-cell} ipython3
 from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis
-from sklearn.metrics import accuracy_score
 
 def get_qda_accuracy(x_train, y_train, x_test, y_test):
     """Fit QDA model to the training data and return the accuracy on the test data."""
     qda = QuadraticDiscriminantAnalysis()
     qda.fit(x_train, y_train)
     y_pred = qda.predict(x_test)
-    accuracy = accuracy_score(y_test, y_pred)
+    accuracy = torch.mean(torch.as_tensor(y_pred == y_test.numpy(), dtype=torch.float))
     return accuracy
 
 model_list = [pca, lda, sqfa_model]