1# Copyright 2017 The TensorFlow Authors. All Rights Reserved.
2#
3# Licensed under the Apache License, Version 2.0 (the "License");
4# you may not use this file except in compliance with the License.
5# You may obtain a copy of the License at
6#
7#     http://www.apache.org/licenses/LICENSE-2.0
8#
9# Unless required by applicable law or agreed to in writing, software
10# distributed under the License is distributed on an "AS IS" BASIS,
11# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12# See the License for the specific language governing permissions and
13# limitations under the License.
14# ==============================================================================
15"""Fashion-MNIST dataset.
16"""
17from __future__ import absolute_import
18from __future__ import division
19from __future__ import print_function
20
21import gzip
22import os
23
24import numpy as np
25
26from tensorflow.python.keras.utils.data_utils import get_file
27from tensorflow.python.util.tf_export import keras_export
28
29
30@keras_export('keras.datasets.fashion_mnist.load_data')
31def load_data():
32  """Loads the Fashion-MNIST dataset.
33
34  This is a dataset of 60,000 28x28 grayscale images of 10 fashion categories,
35  along with a test set of 10,000 images. This dataset can be used as
36  a drop-in replacement for MNIST. The class labels are:
37
38  | Label | Description |
39  |:-----:|-------------|
40  |   0   | T-shirt/top |
41  |   1   | Trouser     |
42  |   2   | Pullover    |
43  |   3   | Dress       |
44  |   4   | Coat        |
45  |   5   | Sandal      |
46  |   6   | Shirt       |
47  |   7   | Sneaker     |
48  |   8   | Bag         |
49  |   9   | Ankle boot  |
50
51  Returns:
52      Tuple of Numpy arrays: `(x_train, y_train), (x_test, y_test)`.
53
54      **x_train, x_test**: uint8 arrays of grayscale image data with shape
55        (num_samples, 28, 28).
56
57      **y_train, y_test**: uint8 arrays of labels (integers in range 0-9)
58        with shape (num_samples,).
59
60  License:
61      The copyright for Fashion-MNIST is held by Zalando SE.
62      Fashion-MNIST is licensed under the [MIT license](
63      https://github.com/zalandoresearch/fashion-mnist/blob/master/LICENSE).
64
65  """
66  dirname = os.path.join('datasets', 'fashion-mnist')
67  base = 'https://storage.googleapis.com/tensorflow/tf-keras-datasets/'
68  files = [
69      'train-labels-idx1-ubyte.gz', 'train-images-idx3-ubyte.gz',
70      't10k-labels-idx1-ubyte.gz', 't10k-images-idx3-ubyte.gz'
71  ]
72
73  paths = []
74  for fname in files:
75    paths.append(get_file(fname, origin=base + fname, cache_subdir=dirname))
76
77  with gzip.open(paths[0], 'rb') as lbpath:
78    y_train = np.frombuffer(lbpath.read(), np.uint8, offset=8)
79
80  with gzip.open(paths[1], 'rb') as imgpath:
81    x_train = np.frombuffer(
82        imgpath.read(), np.uint8, offset=16).reshape(len(y_train), 28, 28)
83
84  with gzip.open(paths[2], 'rb') as lbpath:
85    y_test = np.frombuffer(lbpath.read(), np.uint8, offset=8)
86
87  with gzip.open(paths[3], 'rb') as imgpath:
88    x_test = np.frombuffer(
89        imgpath.read(), np.uint8, offset=16).reshape(len(y_test), 28, 28)
90
91  return (x_train, y_train), (x_test, y_test)
92