Add support for images with other than 3 channels

torchvision/transforms.py

@@ -34,7 +34,7 @@ class ToTensor(object):
         else:
             # handle PIL Image
             img = torch.ByteTensor(torch.ByteStorage.from_buffer(pic.tobytes()))
-            img = img.view(pic.size[1], pic.size[0], 3)
+            img = img.view(pic.size[1], pic.size[0], len(pic.mode))
             # put it from HWC to CHW format
             # yikes, this transpose takes 80% of the loading time/CPU
             img = img.transpose(0, 1).transpose(0, 2).contiguous()