added FashionMNIST dataset (#238)

* added FashionMNIST dataset

* documentation

* fixed formatting

* fixed formatting

README.rst

@@ -43,7 +43,7 @@ Datasets
 
 The following dataset loaders are available:
 
--  `MNIST <#mnist>`__
+-  `MNIST and FashionMNIST <#mnist>`__
 -  `COCO (Captioning and Detection) <#coco>`__
 -  `LSUN Classification <#lsun>`__
 -  `ImageFolder <#imagefolder>`__
@@ -77,6 +77,8 @@ MNIST
 ~~~~~
 ``dset.MNIST(root, train=True, transform=None, target_transform=None, download=False)``
 
+``dset.FashionMNIST(root, train=True, transform=None, target_transform=None, download=False)``
+
 ``root``: root directory of dataset where ``processed/training.pt`` and ``processed/test.pt`` exist
 
 ``train``: ``True`` - use training set, ``False`` - use test set.
@@ -390,32 +392,32 @@ For example:
 Utils
 =====
 
-make\_grid(tensor, nrow=8, padding=2, normalize=False, range=None, scale\_each=False, pad\_value=0)
+``make_grid(tensor, nrow=8, padding=2, normalize=False, range=None, scale_each=False, pad_value=0)``
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Given a 4D mini-batch Tensor of shape (B x C x H x W),
 or a list of images all of the same size,
 makes a grid of images
 
-normalize=True will shift the image to the range (0, 1),
+``normalize=True`` will shift the image to the range (0, 1),
 by subtracting the minimum and dividing by the maximum pixel value.
 
-if range=(min, max) where min and max are numbers, then these numbers are used to
+if ``range=(min, max)`` where ``min`` and ``max`` are numbers, then these numbers are used to
 normalize the image.
 
-scale_each=True will scale each image in the batch of images separately rather than
-computing the (min, max) over all images.
+``scale_each=True`` will scale each image in the batch of images separately rather than
+computing the ``(min, max)`` over all images.
 
-pad_value=<float> sets the value for the padded pixels.
+``pad_value=<float>`` sets the value for the padded pixels.
 
 `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)
+``save_image(tensor, filename, nrow=8, padding=2, normalize=False, range=None, scale_each=False, pad_value=0)``
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Saves a given Tensor into an image file.
 
 If given a mini-batch tensor, will save the tensor as a grid of images.
 
-All options after `filename` are passed through to `make_grid`. Refer to it's documentation for
+All options after ``filename`` are passed through to ``make_grid``. Refer to it's documentation for
 more details

torchvision/datasets/__init__.py

@@ -3,7 +3,7 @@ from .folder import ImageFolder
 from .coco import CocoCaptions, CocoDetection
 from .cifar import CIFAR10, CIFAR100
 from .stl10 import STL10
-from .mnist import MNIST
+from .mnist import MNIST, FashionMNIST
 from .svhn import SVHN
 from .phototour import PhotoTour
 from .fakedata import FakeData
@@ -11,5 +11,5 @@ from .fakedata import FakeData
 __all__ = ('LSUN', 'LSUNClass',
            'ImageFolder', 'FakeData',
            'CocoCaptions', 'CocoDetection',
-           'CIFAR10', 'CIFAR100',
+           'CIFAR10', 'CIFAR100', 'FashionMNIST',
            'MNIST', 'STL10', 'SVHN', 'PhotoTour')

torchvision/datasets/mnist.py

@@ -139,6 +139,17 @@ class MNIST(data.Dataset):
         print('Done!')
 
 
+class FashionMNIST(MNIST):
+    """`Fashion MNIST <https://github.com/zalandoresearch/fashion-mnist>`_ Dataset.
+    """
+    urls = [
+        'http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-images-idx3-ubyte.gz',
+        'http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/train-labels-idx1-ubyte.gz',
+        'http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-images-idx3-ubyte.gz',
+        'http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/t10k-labels-idx1-ubyte.gz',
+    ]
+
+
 def get_int(b):
     return int(codecs.encode(b, 'hex'), 16)