本文整理匯總了Python中tensorflow.python.training.saver.import_meta_graph方法的典型用法代碼示例。如果您正苦於以下問題：Python saver.import_meta_graph方法的具體用法？Python saver.import_meta_graph怎麽用？Python saver.import_meta_graph使用的例子？那麽, 這裏精選的方法代碼示例或許可以為您提供幫助。您也可以進一步了解該方法所在類tensorflow.python.training.saver的用法示例。
在下文中一共展示了saver.import_meta_graph方法的1個代碼示例，這些例子默認根據受歡迎程度排序。您可以為喜歡或者感覺有用的代碼點讚，您的評價將有助於係統推薦出更棒的Python代碼示例。
示例1: load
# 需要導入模塊: from tensorflow.python.training import saver [as 別名]
# 或者: from tensorflow.python.training.saver import import_meta_graph [as 別名]
def load(sess, tags, export_dir):
"""Loads the model from a SavedModel as specified by tags.
Args:
sess: The TensorFlow session to restore the variables.
tags: Set of string tags to identify the required MetaGraphDef. These should
correspond to the tags used when saving the variables using the
SavedModel `save()` API.
export_dir: Directory in which the SavedModel protocol buffer and variables
to be loaded are located.
Returns:
The `MetaGraphDef` protocol buffer loaded in the provided session. This
can be used to further extract signature-defs, collection-defs, etc.
Raises:
RuntimeError: MetaGraphDef associated with the tags cannot be found.
"""
# Build the SavedModel protocol buffer and find the requested meta graph def.
saved_model = _parse_saved_model(export_dir)
found_match = False
for meta_graph_def in saved_model.meta_graphs:
if set(meta_graph_def.meta_info_def.tags) == set(tags):
meta_graph_def_to_load = meta_graph_def
found_match = True
break
if not found_match:
raise RuntimeError("MetaGraphDef associated with tags " + str(tags).strip(
"[]") + " could not be found in SavedModel")
# Build a saver by importing the meta graph def to load.
saver = tf_saver.import_meta_graph(meta_graph_def_to_load)
# Build the checkpoint path where the variables are located.
variables_path = os.path.join(
compat.as_bytes(export_dir),
compat.as_bytes(constants.VARIABLES_DIRECTORY),
compat.as_bytes(constants.VARIABLES_FILENAME))
# Restore the variables using the built saver in the provided session.
saver.restore(sess, variables_path)
# Get asset tensors, if any.
asset_tensors_dictionary = _get_asset_tensors(export_dir,
meta_graph_def_to_load)
# TODO(sukritiramesh): Add support for a single main op to run upon load,
# which will supersede the legacy_init_op.
legacy_init_op_tensor = _get_legacy_init_op_tensor(meta_graph_def_to_load)
if legacy_init_op_tensor is not None:
sess.run(fetches=[legacy_init_op_tensor],
feed_dict=asset_tensors_dictionary)
return meta_graph_def_to_load