From TF 1 X to TF 2.6 (update if you encounter it)

I use it tensorflow 2.6

Generally speaking, in tf 2.x You can use tf.compat.v1.xxxxxx To replace the original tf 1.x Medium tf.xxxxxx

If you make a mistake , Say the function doesn't , In fact, you can go directly to github Of tensorflow Search for this function in the project , see tensorflow How is it used in

Here is a list of the problems encountered

tf 1.x：tf.contrib.layers.xavier_initializer

tf 2.6：tf.keras.initializers.glorot_normal

tf 1.x：tf.contrib.layers

tf 2.6：tf.keras.layers

tf 1.x：from tensorflow.contrib import tpu

tf 2.6：from tensorflow import tpu

tf 1.x：from tensorflow.contrib import data

tf 2.6：from tensorflow import data

tf 1.x：from tensorflow.contrib import metrics

tf 2.6：from tensorflow import metrics

tf 1.x: from tensorflow.contrib import cluster_resolver

tf 2.6: from tensorflow.python.distribute import cluster_resolver

tf 1.x: tf.contrib.layers.layer_norm

tf 2.6: tf.keras.layers.LayerNormalization

The two functions are implemented differently , You can't just change your name

Keras 2.2.5：from keras.optimizers import Adam

Keras 2.6.0：from keras.optimizer_v2.adam import Adam

Keras 2.2.5：from keras.utils import to_categorical

Keras 2.6.0：from keras.utils.np_utils import to_categorical

RuntimeError: When eager execution is enabled, var_list must specify a list or dict of variables to save

When eager execution When it's on ,loss It should be a Python function .
stay Tensorflow 2.0 in ,eager execution Is turned on by default .
therefore , Need to close first eager execution

import tensorflow as tf

tf.compat.v1.disable_eager_execution()