diff --git a/pomegranate/distributions/categorical.py b/pomegranate/distributions/categorical.py
index 93914d44..e5d264e1 100644
--- a/pomegranate/distributions/categorical.py
+++ b/pomegranate/distributions/categorical.py
@@ -175,8 +175,8 @@ def log_probability(self, X):
 		for i in range(self.d):
 			if isinstance(X, torch.masked.MaskedTensor):
 				logp_ = self._log_probs[i][X[:, i]._masked_data]
-				logp_[logp_ == float("-inf")] = 0
-				_inplace_add(logps, logp_)
+				logp_[~X[:, i]._masked_mask] = 0
+				logps += logp_
 			else:
 				logps += self._log_probs[i][X[:, i]]
 
diff --git a/tests/distributions/test_categorical.py b/tests/distributions/test_categorical.py
index 2a3894ca..c9b6fa66 100644
--- a/tests/distributions/test_categorical.py
+++ b/tests/distributions/test_categorical.py
@@ -828,7 +828,8 @@ def test_masked_probability(probs, X, X_masked):
 
 	assert_array_almost_equal(y, d.probability(X_))
 
-	y = [0.042 , 0.0075, 0.001 , 0.003 , 0.007 , 0.001 , 0.014 ]
+	y = [2.1000e-01, 1.5000e-01, 1.0000e+00, 3.0000e-03, 1.0000e-01, 1.0000e-03,
+        1.4000e-01]
 	assert_array_almost_equal(y, d.probability(X_masked))
 
 
@@ -842,7 +843,8 @@ def test_masked_log_probability(probs, X, X_masked):
 
 	assert_array_almost_equal(y, d.log_probability(X_))
 
-	y = numpy.log([0.042 , 0.0075, 0.001 , 0.003 , 0.007 , 0.001 , 0.014 ])
+	y = numpy.log([2.1000e-01, 1.5000e-01, 1.0000e+00, 3.0000e-03, 1.0000e-01, 
+		1.0000e-03, 1.4000e-01])
 	assert_array_almost_equal(y, d.log_probability(X_masked))
 
 def test_masked_summarize(X, X_masked, w):