tensorflow 获取checkpoint中的变量列表实例

方式1：静态获取,通过直接解析checkpoint文件获取变量名及变量值

通过

reader = tf.train.NewCheckpointReader(model_path)

或者通过：

from tensorflow.python import pywrap_tensorflow
reader = pywrap_tensorflow.NewCheckpointReader(model_path)

代码：

model_path = "./checkpoints/model.ckpt-75000"
## 下面两个reader作用等价
#reader = pywrap_tensorflow.NewCheckpointReader(model_path)
reader = tf.train.NewCheckpointReader(model_path)

## 用reader获取变量字典,key是变量名,value是变量的shape
var_to_shape_map = reader.get_variable_to_shape_map()
for var_name in var_to_shape_map.keys():
#用reader获取变量值
var_value = reader.get_tensor(var_name)

print("var_name",var_name)
print("var_value",var_value)

方式2：动态获取,先加载checkpoint模型,然后用graph.get_tensor_by_name()获取变量值

代码 (注意：要先在脚本中构建model中对应的变量及scope):

model_path = "./checkpoints/model.ckpt-75000"
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
with tf.Session(config=config) as sess:
 ## 获取待加载的变量列表
 trainable_vars = tf.trainable_variables()
 g_vars = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,scope="generator")
 d_vars = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,scope='discriminator')
 flow_vars = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,scope='flow_net')
 var_restore = g_vars + d_vars

## 仅加载目标变量
 loader = tf.train.Saver(var_restore)
 loader.restore(sess,model_path)

## 显示加载的变量值
 graph = tf.get_default_graph()
 for var in var_restore:
  tensor = graph.get_tensor_by_name(var.name)
  print("=======变量名=======",tensor)
  print("-------变量值-------",sess.run(tensor))

来源：https://blog.csdn.net/Strive_For_Future/article/details/99675628