fix:error message of to_tensor (#1000)

* fix:error message of to_tensor

The error "pic should be PIL Image or ndarray. Got '<numpy.ndarray>'" is confusing.

* fix:a clearer function name

_is_numpy_image is clearer than _is_numpy_image_dim

* fix:add a test case

Add a test case in test/test_transforms.py to test the error message

* fix:pass ci check

* fix:wrong random matrix

test/test_transforms.py

@@ -434,6 +434,13 @@ class Tester(unittest.TestCase):
         height, width = 4, 4
         trans = transforms.ToTensor()
 
+        with self.assertRaises(TypeError):
+            trans(np.random.rand(1, height, width).tolist())
+
+        with self.assertRaises(ValueError):
+            trans(np.random.rand(height))
+            trans(np.random.rand(1, 1, height, width))
+
         for channels in test_channels:
             input_data = torch.ByteTensor(channels, height, width).random_(0, 255).float().div_(255)
             img = transforms.ToPILImage()(input_data)

torchvision/transforms/functional.py

@@ -31,8 +31,12 @@ def _is_tensor_image(img):
     return torch.is_tensor(img) and img.ndimension() == 3
 
 
+def _is_numpy(img):
+    return isinstance(img, np.ndarray)
+
+
 def _is_numpy_image(img):
-    return isinstance(img, np.ndarray) and (img.ndim in {2, 3})
+    return img.ndim in {2, 3}
 
 
 def to_tensor(pic):
@@ -46,9 +50,12 @@ def to_tensor(pic):
     Returns:
         Tensor: Converted image.
     """
-    if not(_is_pil_image(pic) or _is_numpy_image(pic)):
+    if not(_is_pil_image(pic) or _is_numpy(pic)):
         raise TypeError('pic should be PIL Image or ndarray. Got {}'.format(type(pic)))
 
+    if _is_numpy(pic) and not _is_numpy_image(pic):
+        raise ValueError('pic should be 2/3 dimensional. Got {} dimensions.'.format(pic.ndim))
+
     if isinstance(pic, np.ndarray):
         # handle numpy array
         if pic.ndim == 2: