diff --git a/doctr/models/classification/predictor/pytorch.py b/doctr/models/classification/predictor/pytorch.py
index af59b4a8c7..280743af4d 100644
--- a/doctr/models/classification/predictor/pytorch.py
+++ b/doctr/models/classification/predictor/pytorch.py
@@ -37,6 +37,9 @@ def forward(
         self,
         inputs: list[np.ndarray],
     ) -> list[list[int] | list[float]]:
+        if len(inputs) == 0:
+            return [[], [], []]
+
         # Dimension check
         if any(input.ndim != 3 for input in inputs):
             raise ValueError("incorrect input shape: all inputs are expected to be multi-channel 2D images.")
diff --git a/tests/pytorch/test_models_classification_pt.py b/tests/pytorch/test_models_classification_pt.py
index fc0bbad94d..e5d9add27e 100644
--- a/tests/pytorch/test_models_classification_pt.py
+++ b/tests/pytorch/test_models_classification_pt.py
@@ -137,6 +137,7 @@ def test_crop_orientation_model(mock_text_box):
     # 270 degrees is equivalent to -90 degrees
     assert classifier([text_box_0, text_box_270, text_box_180, text_box_90])[1] == [0, -90, 180, 90]
     assert all(isinstance(pred, float) for pred in classifier([text_box_0, text_box_270, text_box_180, text_box_90])[2])
+    assert classifier([]) == [[], [], []]
 
     # Test with disabled predictor
     classifier = classification.crop_orientation_predictor(
@@ -147,6 +148,7 @@ def test_crop_orientation_model(mock_text_box):
         [0, 0, 0, 0],
         [1.0, 1.0, 1.0, 1.0],
     ]
+    assert classifier([]) == [[], [], []]
 
     # Test custom model loading
     classifier = classification.crop_orientation_predictor(
@@ -182,6 +184,7 @@ def test_page_orientation_model(mock_payslip):
     # 270 degrees is equivalent to -90 degrees
     assert classifier([text_box_0, text_box_270, text_box_180, text_box_90])[1] == [0, -90, 180, 90]
     assert all(isinstance(pred, float) for pred in classifier([text_box_0, text_box_270, text_box_180, text_box_90])[2])
+    assert classifier([]) == [[], [], []]
 
     # Test with disabled predictor
     classifier = classification.page_orientation_predictor(
@@ -192,6 +195,7 @@ def test_page_orientation_model(mock_payslip):
         [0, 0, 0, 0],
         [1.0, 1.0, 1.0, 1.0],
     ]
+    assert classifier([]) == [[], [], []]
 
     # Test custom model loading
     classifier = classification.page_orientation_predictor(