【Tensorflow2】tensorflow1.x-tensorflow2.x一些接口的转变

1. AttributeError: ‘BatchDataset’ object has no attribute ‘make_one_shot_iterator’

tensorflow1.X版本写法：

data_it = dataset.make_one_shot_iterator()

tensorflow2.X版本写法：

data_it = tf.compat.v1.data.make_one_shot_iterator(dataset)

2. AttributeError: module ‘tensorflow_core._api.v2.random’ has no attribute ‘set_random_seed’

tensorflow1.X版本写法：

seed=tf.random.set_random_seed(1234)

tensorflow2.X版本写法：

seed=tf.random.set_seed(1234)

3. AttributeError: module ‘tensorflow’ has no attribute ‘decode_raw’

tensorflow1.X版本写法：

gt_binary_image = tf.decode_raw(features['gt_binary_image_raw'], tf.uint8)

tensorflow2.X版本写法：

gt_binary_image = tf.io.decode_raw(features['gt_binary_image_raw'], tf.uint8)

4. AttributeError: module ‘tensorflow’ has no attribute ‘FixedLenFeature’

tensorflow1.X版本写法：

'gt_image_raw': tf.FixedLenFeature([], tf.string)

tensorflow2.X版本写法：

'gt_image_raw': tf.io.FixedLenFeature([], tf.string)

5. AttributeError: module ‘tensorflow’ has no attribute ‘parse_single_example’

tensorflow1.X版本写法：

features = tf.parse_single_example

tensorflow2.X版本写法：

features = tf.io.parse_single_example(dataset)