Question

我使用带有张量流的keras。我正确地导出模型，但无法创建原型文件。

问题在于我收到错误：

文件＆＃34; keras_indians.py＆＃34;，第103行，in minimal_graph = convert_variables_to_constants（sess，sess.graph_def，[＆＃39; Dense3 / output_node＆＃39;]）文件＆＃34; /usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/graph_util.py" ;, 第234行，在convert_variables_to_constants中 inference_graph = extract_sub_graph（input_graph_def，output_node_names）文件＆＃34; /usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/graph_util.py" ;, 第158行，在extract_sub_graph中在name_to_node_map中断言d，＆＃34;％s不在图表中＃34; ％d AssertionError：Dense3 / output_node不在图中

我尝试了最后一个节点名称的任何可能的变体，但它不起作用任何帮助都非常感谢！

以下是代码：

import tensorflow as tf
from keras.models import Sequential
from keras.layers import Dense

import numpy


# fix random seed for reproducibility
seed = 7
numpy.random.seed(seed)
# load pima indians dataset
dataset = numpy.loadtxt("pima-indians-diabetes.csv", delimiter=",")
# split into input (X) and output (Y) variables
X = dataset[:,0:8]
Y = dataset[:,8]
# create model
model = Sequential()
model.add(Dense(12, input_dim=8, init='uniform', activation='relu'))
model.add(Dense(8, init='uniform', activation='relu'))
model.add(Dense(1, init='uniform', activation='sigmoid', name='output_node'))

# Compile model
model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])
# Fit the model
model.fit(X, Y, nb_epoch=80, batch_size=10)
# evaluate the model


scores = model.evaluate(X, Y)
print("RESULT IS %s: %.2f%%" % (model.metrics_names[1], scores[1]*100))



## Exporting the model ##

print ("Exporting ")


# THIS WORKS !!
#from keras import backend as K
#sess = K.get_session()
#and go about exporting the model as in the tutorial (Note that import for exporter has changed)

#from tensorflow.contrib.session_bundle import exporter

#K.set_learning_phase(0)
#export_path = "./"
#export_version = 1
#saver = tf.train.Saver(sharded=True)
#model_exporter = exporter.Exporter(saver)
#signature = exporter.classification_signature(input_tensor=model.input,
#                                             scores_tensor=model.output)
#model_exporter.init(sess.graph.as_graph_def(),
#                   default_graph_signature=signature)
#model_exporter.export(export_path, tf.constant(export_version), sess)

# END THIS WORKS!


#OPTION 2 FROM: https://blog.keras.io/keras-as-a-simplified-interface-to-tensorflow-tutorial.html

from keras import backend as K

K.set_learning_phase(0)  # all new operations will be in test mode from now on
# added by nir
sess = K.get_session()

# serialize the model and get its weights, for quick re-building
config = model.get_config()
weights = model.get_weights()

# re-build a model where the learning phase is now hard-coded to 0
from keras.models import model_from_config
new_model = model.from_config(config)
new_model.set_weights(weights)


#from tensorflow_serving.session_bundle import exporter
from tensorflow.contrib.session_bundle import exporter

export_path = "./" # where to save the exported graph
export_version = 18 # version number (integer) 

saver = tf.train.Saver(sharded=True)
model_exporter = exporter.Exporter(saver)
signature = exporter.classification_signature(input_tensor=new_model.input,
                                              scores_tensor=new_model.output)
model_exporter.init(sess.graph.as_graph_def(),
                    default_graph_signature=signature)
model_exporter.export(export_path, tf.constant(export_version), sess)


#print  sess.graph.as_graph_def();
print ("Export Done")




## Create proto files for grpc ##

from tensorflow.python.framework.graph_util import convert_variables_to_constants

minimal_graph = convert_variables_to_constants(sess, sess.graph_def, ['Dense3/output_node']) 

tf.train.write_graph(minimal_graph, '.', 'minimal_graph.proto', as_text=False)
tf.train.write_graph(minimal_graph, '.', 'minimal_graph.txt', as_text=True)

Answer 1

直到现在，当我循环遍历图表中创建的所有节点时，我一直在努力解决同样的问题。我发现，在使用下面的代码调试图形后，保存程序用前缀“restore / ***”重写所有节点的名称，找到最后一个节点并在freezeGraph脚本中使用它。

var fs = require('fs');
var path = 'D:\\test2\\2ss\\hhh\\gu';
if(fs.makeDirectory(path))
console.log('"'+path+'" was created.');
else
console.log('"'+path+'" is NOT created.');
phantom.exit();

更多信息：

https://github.com/tensorflow/tensorflow/issues/3986

Tensorflow: What are the "output_node_names" for freeze_graph.py in the model_with_buckets model?

https://www.tensorflow.org/how_tos/tool_developers/

Tensorflow freeze_graph script failing on model defined with Keras

tensorflow - 创建proto文件错误＆＃34;在name_to_node_map中断言d，＆＃34;

1 个答案: