Python cifar10.Cifar10DataSet() Examples

The following are 13 code examples of cifar10.Cifar10DataSet(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module cifar10 , or try the search function .
Example #1
Source File: cifar10_main.py    From DOTA_models with Apache License 2.0 5 votes vote down vote up 
def input_fn(subset, num_shards):
  """Create input graph for model.

  Args:
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  if subset == 'train':
    batch_size = FLAGS.train_batch_size
  elif subset == 'validate' or subset == 'eval':
    batch_size = FLAGS.eval_batch_size
  else:
    raise ValueError('Subset must be one of \'train\', \'validate\' and \'eval\'')
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and FLAGS.use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(FLAGS.data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #2
Source File: cifar10_main.py    From yolo_v2 with Apache License 2.0 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #3
Source File: cifar10_main.py    From Gun-Detector with Apache License 2.0 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #4
Source File: cifar10_main.py    From uai-sdk with Apache License 2.0 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #5
Source File: cifar10_main.py    From hands-detection with MIT License 5 votes vote down vote up 
def input_fn(subset, num_shards):
  """Create input graph for model.

  Args:
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  if subset == 'train':
    batch_size = FLAGS.train_batch_size
  elif subset == 'validate' or subset == 'eval':
    batch_size = FLAGS.eval_batch_size
  else:
    raise ValueError('Subset must be one of \'train\', \'validate\' and \'eval\'')
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and FLAGS.use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(FLAGS.data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #6
Source File: cifar10_myMain.py    From nasbot with MIT License 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #7
Source File: cifar10_main.py    From Live-feed-object-device-identification-using-Tensorflow-and-OpenCV with Apache License 2.0 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #8
Source File: cifar10_main.py    From object_detection_kitti with Apache License 2.0 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #9
Source File: cifar10_myMain.py    From dragonfly with MIT License 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #10
Source File: cifar10_main.py    From object_detection_with_tensorflow with MIT License 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #11
Source File: cifar10_main.py    From g-tensorflow-models with Apache License 2.0 5 votes vote down vote up 
def input_fn(data_dir,
             subset,
             num_shards,
             batch_size,
             use_distortion_for_training=True):
  """Create input graph for model.

  Args:
    data_dir: Directory where TFRecords representing the dataset are located.
    subset: one of 'train', 'validate' and 'eval'.
    num_shards: num of towers participating in data-parallel training.
    batch_size: total batch size for training to be divided by the number of
    shards.
    use_distortion_for_training: True to use distortions.
  Returns:
    two lists of tensors for features and labels, each of num_shards length.
  """
  with tf.device('/cpu:0'):
    use_distortion = subset == 'train' and use_distortion_for_training
    dataset = cifar10.Cifar10DataSet(data_dir, subset, use_distortion)
    image_batch, label_batch = dataset.make_batch(batch_size)
    if num_shards <= 1:
      # No GPU available or only 1 GPU.
      return [image_batch], [label_batch]

    # Note that passing num=batch_size is safe here, even though
    # dataset.batch(batch_size) can, in some cases, return fewer than batch_size
    # examples. This is because it does so only when repeating for a limited
    # number of epochs, but our dataset repeats forever.
    image_batch = tf.unstack(image_batch, num=batch_size, axis=0)
    label_batch = tf.unstack(label_batch, num=batch_size, axis=0)
    feature_shards = [[] for i in range(num_shards)]
    label_shards = [[] for i in range(num_shards)]
    for i in xrange(batch_size):
      idx = i % num_shards
      feature_shards[idx].append(image_batch[i])
      label_shards[idx].append(label_batch[i])
    feature_shards = [tf.parallel_stack(x) for x in feature_shards]
    label_shards = [tf.parallel_stack(x) for x in label_shards]
    return feature_shards, label_shards
Example #12
Source File: cifar10_main.py    From DOTA_models with Apache License 2.0 4 votes vote down vote up 
def main(unused_argv):
  # The env variable is on deprecation path, default is set to off.
  os.environ['TF_SYNC_ON_FINISH'] = '0'

  if FLAGS.num_gpus < 0:
    raise ValueError(
        'Invalid GPU count: \"num_gpus\" must be 0 or a positive integer.')
  if FLAGS.num_gpus == 0 and not FLAGS.is_cpu_ps:
    raise ValueError(
        'No GPU available for use, must use CPU as parameter server.')
  if (FLAGS.num_layers - 2) % 6 != 0:
    raise ValueError('Invalid num_layers parameter.')
  if FLAGS.num_gpus != 0 and FLAGS.train_batch_size % FLAGS.num_gpus != 0:
    raise ValueError('train_batch_size must be multiple of num_gpus.')
  if FLAGS.num_gpus != 0 and FLAGS.eval_batch_size % FLAGS.num_gpus != 0:
    raise ValueError('eval_batch_size must be multiple of num_gpus.')

  num_eval_examples = cifar10.Cifar10DataSet.num_examples_per_epoch('eval')
  if num_eval_examples % FLAGS.eval_batch_size != 0:
    raise ValueError('validation set size must be multiple of eval_batch_size')

  config = tf.estimator.RunConfig()
  sess_config = tf.ConfigProto()
  sess_config.allow_soft_placement = True
  sess_config.log_device_placement = FLAGS.log_device_placement
  sess_config.intra_op_parallelism_threads = FLAGS.num_intra_threads
  sess_config.inter_op_parallelism_threads = FLAGS.num_inter_threads
  sess_config.gpu_options.force_gpu_compatible = FLAGS.force_gpu_compatible
  config = config.replace(session_config=sess_config)

  classifier = tf.estimator.Estimator(
      model_fn=_resnet_model_fn, model_dir=FLAGS.model_dir, config=config)

  tensors_to_log = {'learning_rate': 'learning_rate'}
  logging_hook = tf.train.LoggingTensorHook(
      tensors=tensors_to_log, every_n_iter=100)

  print('Starting to train...')
  classifier.train(
      input_fn=functools.partial(
          input_fn, subset='train', num_shards=FLAGS.num_gpus),
      steps=FLAGS.train_steps,
      hooks=[logging_hook])

  print('Starting to evaluate...')
  eval_results = classifier.evaluate(
      input_fn=functools.partial(
          input_fn, subset='eval', num_shards=FLAGS.num_gpus),
      steps=num_eval_examples // FLAGS.eval_batch_size)
  print(eval_results)
Example #13
Source File: cifar10_main.py    From hands-detection with MIT License 4 votes vote down vote up 
def main(unused_argv):
  # The env variable is on deprecation path, default is set to off.
  os.environ['TF_SYNC_ON_FINISH'] = '0'

  if FLAGS.num_gpus < 0:
    raise ValueError(
        'Invalid GPU count: \"num_gpus\" must be 0 or a positive integer.')
  if FLAGS.num_gpus == 0 and not FLAGS.is_cpu_ps:
    raise ValueError(
        'No GPU available for use, must use CPU as parameter server.')
  if (FLAGS.num_layers - 2) % 6 != 0:
    raise ValueError('Invalid num_layers parameter.')
  if FLAGS.num_gpus != 0 and FLAGS.train_batch_size % FLAGS.num_gpus != 0:
    raise ValueError('train_batch_size must be multiple of num_gpus.')
  if FLAGS.num_gpus != 0 and FLAGS.eval_batch_size % FLAGS.num_gpus != 0:
    raise ValueError('eval_batch_size must be multiple of num_gpus.')

  num_eval_examples = cifar10.Cifar10DataSet.num_examples_per_epoch('eval')
  if num_eval_examples % FLAGS.eval_batch_size != 0:
    raise ValueError('validation set size must be multiple of eval_batch_size')

  config = tf.estimator.RunConfig()
  sess_config = tf.ConfigProto()
  sess_config.allow_soft_placement = True
  sess_config.log_device_placement = FLAGS.log_device_placement
  sess_config.intra_op_parallelism_threads = FLAGS.num_intra_threads
  sess_config.inter_op_parallelism_threads = FLAGS.num_inter_threads
  sess_config.gpu_options.force_gpu_compatible = FLAGS.force_gpu_compatible
  config = config.replace(session_config=sess_config)

  classifier = tf.estimator.Estimator(
      model_fn=_resnet_model_fn, model_dir=FLAGS.model_dir, config=config)

  tensors_to_log = {'learning_rate': 'learning_rate'}
  logging_hook = tf.train.LoggingTensorHook(
      tensors=tensors_to_log, every_n_iter=100)

  print('Starting to train...')
  classifier.train(
      input_fn=functools.partial(
          input_fn, subset='train', num_shards=FLAGS.num_gpus),
      steps=FLAGS.train_steps,
      hooks=[logging_hook])

  print('Starting to evaluate...')
  eval_results = classifier.evaluate(
      input_fn=functools.partial(
          input_fn, subset='eval', num_shards=FLAGS.num_gpus),
      steps=num_eval_examples // FLAGS.eval_batch_size)
  print(eval_results)