python类get()的实例源码

def __init__(self,
               num,
               batch_size=1,
               progress_bar=False,
               log_epoch=10,
               get_fn=None,
               cycle=False,
               shuffle=True,
               stagnant=False,
               seed=2,
               num_batches=-1):
    """Construct a batch iterator.

    Args:
        data: numpy.ndarray, (N, D), N is the number of examples, D is the
        feature dimension.
        labels: numpy.ndarray, (N), N is the number of examples.
        batch_size: int, batch size.
    """

    self._num = num
    self._batch_size = batch_size
    self._step = 0
    self._num_steps = int(np.ceil(self._num / float(batch_size)))
    if num_batches > 0:
      self._num_steps = min(self._num_steps, num_batches)
    self._pb = None
    self._variables = None
    self._get_fn = get_fn
    self.get_fn = get_fn
    self._cycle = cycle
    self._shuffle_idx = np.arange(self._num)
    self._shuffle = shuffle
    self._random = np.random.RandomState(seed)
    if shuffle:
      self._random.shuffle(self._shuffle_idx)
    self._shuffle_flag = False
    self._stagnant = stagnant
    self._log_epoch = log_epoch
    self._log = logger.get()
    self._epoch = 0
    if progress_bar:
      self._pb = pb.get(self._num_steps)
      pass
    self._mutex = threading.Lock()
    pass

def __init__(self,
               num,
               batch_size=1,
               progress_bar=False,
               log_epoch=10,
               get_fn=None,
               cycle=False,
               shuffle=True,
               stagnant=False,
               seed=2,
               num_batches=-1):
    """Construct a batch iterator.

    Args:
        data: numpy.ndarray, (N, D), N is the number of examples, D is the
        feature dimension.
        labels: numpy.ndarray, (N), N is the number of examples.
        batch_size: int, batch size.
    """

    self._num = num
    self._batch_size = batch_size
    self._step = 0
    self._num_steps = int(np.ceil(self._num / float(batch_size)))
    if num_batches > 0:
      self._num_steps = min(self._num_steps, num_batches)
    self._pb = None
    self._variables = None
    self._get_fn = get_fn
    self.get_fn = get_fn
    self._cycle = cycle
    self._shuffle_idx = np.arange(self._num)
    self._shuffle = shuffle
    self._random = np.random.RandomState(seed)
    if shuffle:
      self._random.shuffle(self._shuffle_idx)
    self._shuffle_flag = False
    self._stagnant = stagnant
    self._log_epoch = log_epoch
    self._log = logger.get()
    self._epoch = 0
    if progress_bar:
      self._pb = pb.get(self._num_steps)
      pass
    self._mutex = threading.Lock()
    pass

def __init__(self,
               num,
               batch_size=1,
               progress_bar=False,
               log_epoch=10,
               get_fn=None,
               cycle=False,
               shuffle=True,
               stagnant=False,
               seed=2,
               num_batches=-1):
    """Construct a batch iterator.

    Args:
        data: numpy.ndarray, (N, D), N is the number of examples, D is the
        feature dimension.
        labels: numpy.ndarray, (N), N is the number of examples.
        batch_size: int, batch size.
    """

    self._num = num
    self._batch_size = batch_size
    self._step = 0
    self._num_steps = int(np.ceil(self._num / float(batch_size)))
    if num_batches > 0:
      self._num_steps = min(self._num_steps, num_batches)
    self._pb = None
    self._variables = None
    self._get_fn = get_fn
    self.get_fn = get_fn
    self._cycle = cycle
    self._shuffle_idx = np.arange(self._num)
    self._shuffle = shuffle
    self._random = np.random.RandomState(seed)
    if shuffle:
      self._random.shuffle(self._shuffle_idx)
    self._shuffle_flag = False
    self._stagnant = stagnant
    self._log_epoch = log_epoch
    self._log = logger.get()
    self._epoch = 0
    if progress_bar:
      self._pb = pb.get(self._num_steps)
      pass
    self._mutex = threading.Lock()
    pass