Permalink
Cannot retrieve contributors at this time
Fetching contributors…
Cannot retrieve contributors at this time
| # Copyright 2017 The TensorFlow Authors. All Rights Reserved. | |
| # | |
| # Licensed under the Apache License, Version 2.0 (the "License"); | |
| # you may not use this file except in compliance with the License. | |
| # You may obtain a copy of the License at | |
| # | |
| # http://www.apache.org/licenses/LICENSE-2.0 | |
| # | |
| # Unless required by applicable law or agreed to in writing, software | |
| # distributed under the License is distributed on an "AS IS" BASIS, | |
| # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
| # See the License for the specific language governing permissions and | |
| # limitations under the License. | |
| # ============================================================================== | |
| """CIFAR-10 data set. | |
| See http://www.cs.toronto.edu/~kriz/cifar.html. | |
| """ | |
| import os | |
| import tensorflow as tf | |
| flags = tf.app.flags | |
| FLAGS = flags.FLAGS | |
| flags.DEFINE_bool('random_brightness', False, '') | |
| flags.DEFINE_bool('random_contrast', False, '') | |
| flags.DEFINE_bool('return_dict', False, '') | |
| HEIGHT = 224 | |
| WIDTH = 224 | |
| DEPTH = 3 | |
| class Dataset(object): | |
| def __init__(self, data_dir, subset='train', use_distortion=True): | |
| self.data_dir = data_dir | |
| self.subset = subset | |
| self.use_distortion = use_distortion | |
| def get_filenames(self): | |
| if self.subset in ['train', 'valid', 'test']: | |
| return [os.path.join(self.data_dir, self.subset + '.tfrecords')] | |
| else: | |
| raise ValueError('Invalid data subset "%s"' % self.subset) | |
| def parser(self, serialized_example): | |
| """Parses a single tf.Example into image and label tensors.""" | |
| # Dimensions of the images in the CIFAR-10 dataset. | |
| # See http://www.cs.toronto.edu/~kriz/cifar.html for a description of the | |
| # input format. | |
| features = tf.parse_single_example( | |
| serialized_example, | |
| features={ | |
| 'id': tf.FixedLenFeature([], tf.string), | |
| 'image': tf.FixedLenFeature([], tf.string), | |
| 'label': tf.FixedLenFeature([], tf.int64), | |
| }) | |
| image = features['image'] | |
| #image = tf.expand_dims(image, axis=-1) | |
| print('--------------1', image) | |
| #image = tf.decode_raw(image, tf.uint8) | |
| image = tf.image.decode_png(image) | |
| #image.set_shape([HEIGHT * WIDTH * DEPTH]) | |
| print('--------------2', image, image.shape) | |
| #image = tf.cast( | |
| # tf.reshape(image, [HEIGHT, WIDTH, DEPTH]), | |
| # tf.float32) | |
| image = tf.cast(image, tf.float32) | |
| label = tf.cast(features['label'], tf.int32) | |
| id = features['id'] | |
| # Custom preprocessing. | |
| image = self.preprocess(image) | |
| # if not FLAGS.return_dict: | |
| # return id, image, label | |
| # else: | |
| # return {'id': id, 'image': image}, label | |
| return image, label | |
| def make_batch(self, batch_size, filenames=None, repeat=None, initializable=None): | |
| """Read the images and labels from 'filenames'.""" | |
| filenames = filenames or self.get_filenames() | |
| if initializable is None: | |
| initializable = self.subset != 'train' | |
| # Repeat infinitely. | |
| dataset = tf.data.TFRecordDataset(filenames).repeat() | |
| # Parse records. | |
| dataset = dataset.map( | |
| self.parser, num_parallel_calls=batch_size) | |
| # Potentially shuffle records. | |
| if self.subset == 'train': | |
| min_queue_examples = int( | |
| Dataset.num_examples_per_epoch(self.subset) * 0.4) | |
| # Ensure that the capacity is sufficiently large to provide good random | |
| # shuffling. | |
| dataset = dataset.shuffle(buffer_size=min_queue_examples + 3 * batch_size) | |
| # Batch it up. | |
| dataset = dataset.batch(batch_size) | |
| assert tf.executing_eagerly() | |
| return dataset | |
| def preprocess(self, image): | |
| """Preprocess a single image in [height, width, depth] layout.""" | |
| if self.subset == 'train' and self.use_distortion: | |
| # Pad 4 pixels on each dimension of feature map, done in mini-batch | |
| #... yes should do something like below.. but you will see with dataset.map.. not ok as summary without scope and finally graph has no these summaries | |
| # refer to https://stackoverflow.com/questions/47345394/image-summaries-with-tensorflows-dataset-api TODO FIXME | |
| tf.summary.image('image', image) | |
| #print('--------', image) | |
| #image = tf.image.resize_image_with_crop_or_pad(image, 256, 256) | |
| #image = tf.image.resize_images(image, [256, 256], method=0) | |
| #image = tf.random_crop(image, [HEIGHT, WIDTH, DEPTH]) | |
| image = tf.image.random_flip_left_right(image) | |
| if FLAGS.random_brightness: | |
| image = tf.image.random_brightness(image, | |
| max_delta=63) | |
| if FLAGS.random_contrast: | |
| distorted_image = tf.image.random_contrast(image, | |
| lower=0.2, | |
| upper=1.8) | |
| tf.summary.image('image/distort', image) | |
| else: | |
| image = tf.image.resize_images(image, [HEIGHT, WIDTH], method=0) | |
| image = tf.cast( | |
| tf.reshape(image, [HEIGHT, WIDTH, DEPTH]), | |
| tf.float32) | |
| return image | |
| @staticmethod | |
| def num_examples_per_epoch(subset='train'): | |
| if subset == 'train': | |
| return 3295 | |
| elif subset == 'valid': | |
| return 367 | |
| elif subset == 'test': | |
| return 1928 | |
| else: | |
| raise ValueError('Invalid data subset "%s"' % subset) |