diff --git a/torchvision/models/detection/retinanet.py b/torchvision/models/detection/retinanet.py
index 446489db989..8bbade08663 100644
--- a/torchvision/models/detection/retinanet.py
+++ b/torchvision/models/detection/retinanet.py
@@ -454,7 +454,7 @@ def postprocess_detections(self, head_outputs, anchors, image_shapes):
 
         return detections
 
-    def _anchors_per_level(self, features, HWA):
+    def _anchors_per_level(self, features: List[Tensor], HWA: int):
         # recover level sizes
         num_anchors_per_level = [x.size(2) * x.size(3) for x in features]
         HW = 0
diff --git a/torchvision/models/detection/ssd.py b/torchvision/models/detection/ssd.py
index 0e70b840a50..12186447796 100644
--- a/torchvision/models/detection/ssd.py
+++ b/torchvision/models/detection/ssd.py
@@ -141,7 +141,7 @@ def __init__(self, backbone: nn.Module, num_classes: int,
         # used only on torchscript mode
         self._has_warned = False
 
-    def _anchors_per_level(self, features, HWA):
+    def _anchors_per_level(self, features: List[Tensor], HWA: int):
         # TODO: Discuss/refactor this workaround
         num_anchors_per_level = [x.size(2) * x.size(3) * anchors for x, anchors in zip(features, self.num_anchors)]
         HW = 0