diff --git a/tests/beignet/features/conftest.py b/tests/beignet/features/conftest.py
deleted file mode 100644
index 72e2b5edd5..0000000000
--- a/tests/beignet/features/conftest.py
+++ /dev/null
@@ -1,8 +0,0 @@
-import pytest
-import torch
-from beignet.features._feature import Feature
-
-
-@pytest.fixture
-def feature() -> Feature:
-    return Feature(torch.tensor([1, 2, 3]))
diff --git a/tests/beignet/features/test__feature.py b/tests/beignet/features/test__feature.py
index 02d30af4ac..6262f9089d 100644
--- a/tests/beignet/features/test__feature.py
+++ b/tests/beignet/features/test__feature.py
@@ -6,18 +6,15 @@
 
 
 class TestFeature:
-    def test__to_tensor(self, feature: Feature):
-        result = Feature._to_tensor([1, 2, 3])
+    def test___deepcopy__(self):
+        feature = Feature(torch.tensor([1, 2, 3]))
 
-        assert torch.is_tensor(result)
-
-        assert not result.requires_grad
-
-    def test_wrap_like(self):
         with pytest.raises(NotImplementedError):
-            Feature.wrap_like(None, None)
+            copy.deepcopy(feature)
+
+    def test___torch_function__(self):
+        feature = Feature(torch.tensor([1, 2, 3]))
 
-    def test___torch_function__(self, feature: Feature):
         result = feature.__torch_function__(
             torch.add,
             (Feature, torch.Tensor),
@@ -26,18 +23,29 @@ def test___torch_function__(self, feature: Feature):
 
         assert not isinstance(result, Feature)
 
-    def test_device(self, feature: Feature):
+    def test__to_tensor(self):
+        feature = Feature._to_tensor([1, 2, 3])
+
+        assert torch.is_tensor(feature)
+
+        assert not feature.requires_grad
+
+    def test_device(self):
+        feature = Feature(torch.tensor([1, 2, 3]))
+
         assert feature.device == feature.device
 
-    def test_ndim(self, feature: Feature):
-        assert feature.ndim == 1
+    def test_dtype(self):
+        assert Feature(torch.tensor([1, 2, 3])).dtype == torch.int64
 
-    def test_dtype(self, feature: Feature):
-        assert feature.dtype == torch.int64
+    def test_ndim(self):
+        assert Feature(torch.tensor([1, 2, 3])).ndim == 1
 
-    def test_shape(self, feature: Feature):
-        assert feature.shape == (3,)
+    def test_shape(self):
+        assert Feature(torch.tensor([1, 2, 3])).shape == (3,)
 
-    def test___deepcopy__(self, feature: Feature):
+    def test_wrap_like(self):
         with pytest.raises(NotImplementedError):
-            copy.deepcopy(feature)
+            feature = Feature(torch.tensor([1, 2, 3]))
+
+            Feature.wrap_like(feature, torch.tensor([1, 2, 3]))