AttributeError: type object 'DatasetV2' has no attribute 'parallel_interleave'

Question

我在 tensorflow 2 数据集上尝试使用 apply 方法时遇到此错误，这是我编写的代码：

from tensorflow import data as tfdata

dataset = tfdata.Dataset.list_files(
    self._file_pattern,
    shuffle=self._is_training,
    seed=tf.random.set_random_seed(int(time.time() * 1e9)))
if self._is_training:
  dataset = dataset.repeat()

然后应用方法：

# Prefetch data from files.
def _prefetch_dataset(filename):
  dataset = tfdata.TFRecordDataset(
      filename, buffer_size=BUFFER_SIZE).prefetch(1)
  return dataset

dataset = dataset.apply(
    tfdata.Dataset.parallel_interleave(
    _prefetch_dataset, cycle_length=32, sloppy=self._is_training))

Answer 1

在 tensorflow 2.x 中我们不再需要使用 apply 方法。只需使用 dataset = dataset.interleave() 而不是 dataset.apply(tf.data.experimental.parallel_interleave())

AttributeError: type object 'DatasetV2' has no attribute 'parallel_interleave'

AttributeError: type object 'DatasetV2' has no attribute 'parallel_interleave'

python

tensorflow

tensorflow2.0