Merge branch 'master' of github.com:pytorch/vision

Author: chsasank

README.rst

@@ -280,7 +280,7 @@ to be atleast 224.
 The images have to be loaded in to a range of [0, 1] and then
 normalized using `mean=[0.485, 0.456, 0.406]` and `std=[0.229, 0.224, 0.225]`
 
-An example of such normalization can be found in `the imagenet example here` <https://github.com/pytorch/examples/blob/42e5b996718797e45c46a25c55b031e6768f8440/imagenet/main.py#L89-L101>
+An example of such normalization can be found in the imagenet example `here <https://github.com/pytorch/examples/blob/42e5b996718797e45c46a25c55b031e6768f8440/imagenet/main.py#L89-L101>`__
 
 Transforms
 ==========
@@ -410,7 +410,7 @@ computing the ``(min, max)`` over all images.
 
 ``pad_value=<float>`` sets the value for the padded pixels.
 
-`Example usage is given in this notebook` <https://gist.github.com/anonymous/bf16430f7750c023141c562f3e9f2a91>
+Example usage is given in this `notebook <https://gist.github.com/anonymous/bf16430f7750c023141c562f3e9f2a91>`__ 
 
 ``save_image(tensor, filename, nrow=8, padding=2, normalize=False, range=None, scale_each=False, pad_value=0)``
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

setup.py

@@ -11,7 +11,7 @@
 
 requirements = [
     'numpy',
-    'pillow',
+    'pillow >= 4.1.1',
     'six',
     'torch',
 ]

test/test_transforms.py

@@ -86,15 +86,15 @@ def test_scale(self):
         owidth = random.randint(5, 12) * 2
         result = transforms.Compose([
             transforms.ToPILImage(),
-            transforms.Scale((owidth, oheight)),
+            transforms.Scale((oheight, owidth)),
             transforms.ToTensor(),
         ])(img)
         assert result.size(1) == oheight
         assert result.size(2) == owidth
 
         result = transforms.Compose([
             transforms.ToPILImage(),
-            transforms.Scale([owidth, oheight]),
+            transforms.Scale([oheight, owidth]),
             transforms.ToTensor(),
         ])(img)
         assert result.size(1) == oheight
@@ -150,7 +150,7 @@ def test_pad_with_tuple_of_pad_values(self):
         assert output.size[0] == width + padding[0] + padding[2]
         assert output.size[1] == height + padding[1] + padding[3]
 
-    def test_pad_raises_with_invalide_pad_sequence_len(self):
+    def test_pad_raises_with_invalid_pad_sequence_len(self):
         with self.assertRaises(ValueError):
             transforms.Pad(())
 
@@ -264,6 +264,22 @@ def test_tensor_gray_to_pil_image(self):
         assert np.allclose(img_data_short.numpy(), to_tensor(img_short).numpy())
         assert np.allclose(img_data_int.numpy(), to_tensor(img_int).numpy())
 
+    def test_tensor_rgba_to_pil_image(self):
+        trans = transforms.ToPILImage()
+        to_tensor = transforms.ToTensor()
+
+        img_data = torch.Tensor(4, 4, 4).uniform_()
+        img = trans(img_data)
+        assert img.mode == 'RGBA'
+        assert img.getbands() == ('R', 'G', 'B', 'A')
+        r, g, b, a = img.split()
+
+        expected_output = img_data.mul(255).int().float().div(255)
+        assert np.allclose(expected_output[0].numpy(), to_tensor(r).numpy())
+        assert np.allclose(expected_output[1].numpy(), to_tensor(g).numpy())
+        assert np.allclose(expected_output[2].numpy(), to_tensor(b).numpy())
+        assert np.allclose(expected_output[3].numpy(), to_tensor(a).numpy())
+
     def test_ndarray_to_pil_image(self):
         trans = transforms.ToPILImage()
         img_data = torch.ByteTensor(4, 4, 3).random_(0, 255).numpy()

test/test_utils.py

@@ -0,0 +1,31 @@
+import torch
+import torchvision.utils as utils
+import unittest
+
+
+class Tester(unittest.TestCase):
+
+    def test_make_grid_not_inplace(self):
+        t = torch.rand(5, 3, 10, 10)
+        t_clone = t.clone()
+
+        utils.make_grid(t, normalize=False)
+        assert torch.equal(t, t_clone), 'make_grid modified tensor in-place'
+
+        utils.make_grid(t, normalize=True, scale_each=False)
+        assert torch.equal(t, t_clone), 'make_grid modified tensor in-place'
+
+        utils.make_grid(t, normalize=True, scale_each=True)
+        assert torch.equal(t, t_clone), 'make_grid modified tensor in-place'
+
+    def test_make_grid_raises_with_variable(self):
+        t = torch.autograd.Variable(torch.rand(3, 10, 10))
+        with self.assertRaises(TypeError):
+            utils.make_grid(t)
+
+        with self.assertRaises(TypeError):
+            utils.make_grid([t, t, t, t])
+
+
+if __name__ == '__main__':
+    unittest.main()

torchvision/datasets/utils.py

@@ -10,7 +10,7 @@ def check_integrity(fpath, md5):
     md5o = hashlib.md5()
     with open(fpath, 'rb') as f:
         # read in 1MB chunks
-        for chunk in iter(lambda: f.read(1024 * 1024 * 1024), b''):
+        for chunk in iter(lambda: f.read(1024 * 1024), b''):
             md5o.update(chunk)
     md5c = md5o.hexdigest()
     if md5c != md5:

torchvision/models/densenet.py

@@ -8,10 +8,10 @@
 
 
 model_urls = {
-    'densenet121': 'https://download.pytorch.org/models/densenet121-241335ed.pth',
-    'densenet169': 'https://download.pytorch.org/models/densenet169-6f0f7f60.pth',
-    'densenet201': 'https://download.pytorch.org/models/densenet201-4c113574.pth',
-    'densenet161': 'https://download.pytorch.org/models/densenet161-17b70270.pth',
+    'densenet121': 'https://download.pytorch.org/models/densenet121-a639ec97.pth',
+    'densenet169': 'https://download.pytorch.org/models/densenet169-b2777c0a.pth',
+    'densenet201': 'https://download.pytorch.org/models/densenet201-c1103571.pth',
+    'densenet161': 'https://download.pytorch.org/models/densenet161-8d451a50.pth',
 }
 
 

torchvision/transforms.py

@@ -109,6 +109,9 @@ def to_pil_image(pic):
             mode = 'I'
         elif npimg.dtype == np.float32:
             mode = 'F'
+    elif npimg.shape[2] == 4:
+            if npimg.dtype == np.uint8:
+                mode = 'RGBA'
     else:
         if npimg.dtype == np.uint8:
             mode = 'RGB'

torchvision/utils.py

@@ -10,8 +10,8 @@ def make_grid(tensor, nrow=8, padding=2,
     Args:
         tensor (Tensor or list): 4D mini-batch Tensor of shape (B x C x H x W)
             or a list of images all of the same size.
-        nrow (int, optional): Number of rows in grid. Final grid size is
-            (B / nrow, nrow). Default is 8.
+        nrow (int, optional): Number of images displayed in each row of the grid.
+            The Final grid size is (B / nrow, nrow). Default is 8.
         padding (int, optional): amount of padding. Default is 2.
         normalize (bool, optional): If True, shift the image to the range (0, 1),
             by subtracting the minimum and dividing by the maximum pixel value.
@@ -26,14 +26,13 @@ def make_grid(tensor, nrow=8, padding=2,
         See this notebook `here <https://gist.github.com/anonymous/bf16430f7750c023141c562f3e9f2a91>`_
 
     """
+    if not (torch.is_tensor(tensor) or
+            (isinstance(tensor, list) and all(torch.is_tensor(t) for t in tensor))):
+        raise TypeError('tensor or list of tensors expected, got {}'.format(type(tensor)))
+
     # if list of tensors, convert to a 4D mini-batch Tensor
     if isinstance(tensor, list):
-        tensorlist = tensor
-        numImages = len(tensorlist)
-        size = torch.Size(torch.Size([numImages]) + tensorlist[0].size())
-        tensor = tensorlist[0].new(size)
-        for i in irange(numImages):
-            tensor[i].copy_(tensorlist[i])
+        tensor = torch.stack(tensor, dim=0)
 
     if tensor.dim() == 2:  # single image H x W
         tensor = tensor.view(1, tensor.size(0), tensor.size(1))
@@ -45,6 +44,7 @@ def make_grid(tensor, nrow=8, padding=2,
         tensor = torch.cat((tensor, tensor, tensor), 1)
 
     if normalize is True:
+        tensor = tensor.clone()  # avoid modifying tensor in-place
         if range is not None:
             assert isinstance(range, tuple), \
                 "range has to be a tuple (min, max) if specified. min and max are numbers"
@@ -70,14 +70,14 @@ def norm_range(t, range):
     xmaps = min(nrow, nmaps)
     ymaps = int(math.ceil(float(nmaps) / xmaps))
     height, width = int(tensor.size(2) + padding), int(tensor.size(3) + padding)
-    grid = tensor.new(3, height * ymaps + 1 + padding // 2, width * xmaps + 1 + padding // 2).fill_(pad_value)
+    grid = tensor.new(3, height * ymaps + padding, width * xmaps + padding).fill_(pad_value)
     k = 0
     for y in irange(ymaps):
         for x in irange(xmaps):
             if k >= nmaps:
                 break
-            grid.narrow(1, y * height + 1 + padding // 2, height - padding)\
-                .narrow(2, x * width + 1 + padding // 2, width - padding)\
+            grid.narrow(1, y * height + padding, height - padding)\
+                .narrow(2, x * width + padding, width - padding)\
                 .copy_(tensor[k])
             k = k + 1
     return grid