diff --git a/sahi/utils/torch.py b/sahi/utils/torch.py
index a5d289678..74a60270f 100644
--- a/sahi/utils/torch.py
+++ b/sahi/utils/torch.py
@@ -15,37 +15,37 @@ def empty_cuda_cache():
         raise RuntimeError("CUDA not available.")
 
 
-if is_available("torch"):
-
-    @check_requirements(["torch"])
-    def to_float_tensor(img):
-        """
-        Converts a PIL.Image (RGB) or numpy.ndarray (H x W x C) in the range
-        [0, 255] to a torch.FloatTensor of shape (C x H x W).
-        Args:
-            img: np.ndarray
-        Returns:
-            torch.tensor
-        """
-        import torch
+@check_requirements(["torch"])
+def to_float_tensor(img):
+    """
+    Converts a PIL.Image (RGB) or numpy.ndarray (H x W x C) in the range
+    [0, 255] to a torch.FloatTensor of shape (C x H x W).
+    Args:
+        img: np.ndarray
+    Returns:
+        torch.tensor
+    """
+    import torch
 
-        img = img.transpose((2, 0, 1))
-        img = torch.from_numpy(img).float()
-        if img.max() > 1:
-            img /= 255
+    img = img.transpose((2, 0, 1))
+    img = torch.from_numpy(img).float()
+    if img.max() > 1:
+        img /= 255
 
-        return img
+    return img
 
-    @check_requirements(["torch"])
-    def torch_to_numpy(img):
-        import torch
 
-        img = img.numpy()
-        if img.max() > 1:
-            img /= 255
-        return img.transpose((1, 2, 0))
+@check_requirements(["torch"])
+def torch_to_numpy(img):
+    import torch
+
+    img = img.numpy()
+    if img.max() > 1:
+        img /= 255
+    return img.transpose((1, 2, 0))
 
 
+@check_requirements(["torch"])
 def is_torch_cuda_available():
     if is_available("torch"):
         import torch
diff --git a/tests/test_detectron2.py b/tests/test_detectron2.py
index ce48a53a0..13230fae5 100644
--- a/tests/test_detectron2.py
+++ b/tests/test_detectron2.py
@@ -6,7 +6,7 @@
 from sahi.model import Detectron2DetectionModel
 from sahi.utils.cv import read_image
 from sahi.utils.detectron2 import Detectron2TestConstants
-from sahi.utils.import_utils import _torch_version
+from sahi.utils.import_utils import get_package_info
 
 MODEL_DEVICE = "cpu"
 CONFIDENCE_THRESHOLD = 0.5
@@ -14,7 +14,7 @@
 
 # note that detectron2 binaries are available only for linux
 
-if _torch_version == "1.10.2":
+if get_package_info("torch", verbose=False)[1] == "1.10.2":
 
     class TestDetectron2DetectionModel(unittest.TestCase):
         def test_load_model(self):