a Sic>M@s\dZddlZddlmmZddlmZddlm Z ddl m Z ddl m Z ddl mZddl mZdd lmZdd lmZdd lmZdd lmZed edZededZededZdZedgdd6ddZddZ ddZ!ddZ"dd Z#d!d"Z$d#d$Z%d%d&Z&d'd(Z'ed)gdd7d*d+Z(d,d-Z)d.d/Z*d0d1Z+d2d3Z,d4d5Z-dS)8z8Deprecated experimental Keras SavedModel implementation.N)backend) optimizer_v1) optimizer_v2) model_config) saving_utils)utils_v1) mode_keys) LazyLoader) tf_logging) keras_export metrics_libz keras.metrics models_libz keras.models sequentialzkeras.engine.sequentialzsaved_model.jsonz%keras.experimental.export_saved_model)v1FcCsntjddd|r.tjj||t||dnt|||||zt||Wnt yht dYn0dS)a Exports a `tf.keras.Model` as a Tensorflow SavedModel. Note that at this time, subclassed models can only be saved using `serving_only=True`. The exported `SavedModel` is a standalone serialization of Tensorflow objects, and is supported by TF language APIs and the Tensorflow Serving system. To load the model, use the function `tf.compat.v1.keras.experimental.load_from_saved_model`. The `SavedModel` contains: 1. a checkpoint containing the model weights. 2. a `SavedModel` proto containing the Tensorflow backend graph. Separate graphs are saved for prediction (serving), train, and evaluation. If the model has not been compiled, then only the graph computing predictions will be exported. 3. the model's json config. If the model is subclassed, this will only be included if the model's `get_config()` method is overwritten. Example: ```python import tensorflow as tf # Create a tf.keras model. model = tf.keras.Sequential() model.add(tf.keras.layers.Dense(1, input_shape=[10])) model.summary() # Save the tf.keras model in the SavedModel format. path = '/tmp/simple_keras_model' tf.compat.v1.keras.experimental.export_saved_model(model, path) # Load the saved keras model back. new_model = tf.compat.v1.keras.experimental.load_from_saved_model(path) new_model.summary() ``` Args: model: A `tf.keras.Model` to be saved. If the model is subclassed, the flag `serving_only` must be set to True. saved_model_path: a string specifying the path to the SavedModel directory. custom_objects: Optional dictionary mapping string names to custom classes or functions (e.g. custom loss functions). as_text: bool, `False` by default. Whether to write the `SavedModel` proto in text format. Currently unavailable in serving-only mode. input_signature: A possibly nested sequence of `tf.TensorSpec` objects, used to specify the expected model inputs. See `tf.function` for more details. serving_only: bool, `False` by default. When this is true, only the prediction graph is saved. Raises: NotImplementedError: If the model is a subclassed model, and serving_only is False. ValueError: If the input signature cannot be inferred from the model. AssertionError: If the SavedModel directory already exists and isn't empty. z`tf.keras.experimental.export_saved_model` is deprecatedand will be removed in a future version. Please use `model.save(..., save_format="tf")` or `tf.keras.models.save_model(..., save_format="tf")`. stacklevel) signatureszOSkipped saving model JSON, subclassed model does not have get_config() defined.N) warningswarntf saved_modelsavertrace_model_call_save_v1_format_export_model_jsonNotImplementedErrorloggingwarning)modelsaved_model_pathcustom_objectsas_textinput_signature serving_onlyr%a/var/www/html/django/DPS/env/lib/python3.9/site-packages/keras/saving/saved_model_experimental.pyexport_saved_model1s&F   r'cCsb|}tjjt|tjt}tjj |d}| |Wdn1sT0YdS)z?Saves model configuration as a json string under assets folder.wN) to_jsonriogfilejoin_get_or_create_assets_dircompatr"SAVED_MODEL_FILENAME_JSONGFilewrite)rr model_jsonmodel_json_filepathfr%r%r&rs rcCs$t|t|}|j|ddd|S)z@Saves model weights in checkpoint format under variables folder.rT save_format overwrite)_get_or_create_variables_dir_get_variables_path save_weights)rr checkpoint_prefixr%r%r&_export_model_variablessr<c Cs|js*t|tjr"|js*tdntdtjj |}t ||}|||||d}d}|j rt|j t jtjfrttjj|fi|d}ttjj|fi|n tdttjj|fi|||dS)z&Exports model to v1 SavedModel format.zWeights for sequential model have not yet been created. Weights are created when the Model is first called on inputs or `build()` is called with an `input_shape`, or the first layer in the model has `input_shape` during construction.zZSubclassed models can only be exported for serving. Please set argument serving_only=True.)builderrr!checkpoint_pathr#FTzModel was compiled with an optimizer, but the optimizer is not from `tf.train` (e.g. `tf.train.AdagradOptimizer`). Only the serving graph was exported. The train and evaluate graphs were not added to the SavedModel.N)_is_graph_network isinstancer Sequentialbuilt ValueErrorrr __internal__rSavedModelBuilderr< optimizerr TFOptimizerr OptimizerV2 _export_moderModeKeysTRAINTESTrrPREDICTr) rpathr!r"r#r=r>Z export_argshas_saved_varsr%r%r&rsF    rcCstjj|\}}}|S)z?Returns list of all checkpointed saveable objects in the model.)rrDtrackingObjectGraphViewserialize_object_graph)rvar_list_r%r%r& _get_var_listsrUcCstj|j|j|jdS)N)shapedtypename)r placeholderrVrWrX)specr%r%r&create_placeholders r[c Cs|tjjk}|r.|js.td|jd|dtjj }t } t |tjjk||durnd} ntjt|} tj|| ||d} |r| tjjjj| jjd} |tjjkr| | jj} n|tjjkr| n| | tjjjj | j!tjj" t#| } |r8| $|}|%n.t&||| | | $|| j'|dddd|_(|j)t*j+|t,| |tjjj-j.| dd tjj/| d Wdn1s0YWdWddS1s0YWdn1s0YdS) aExports a model, and optionally saves new vars from the clone model. Args: mode: A `tf.estimator.ModeKeys` string. has_saved_vars: A `boolean` indicating whether the SavedModel has already exported variables. builder: A `SavedModelBuilder` object. model: A `tf.keras.Model` object. custom_objects: A dictionary mapping string names to custom classes or functions. checkpoint_path: String path to checkpoint. input_signature: Nested TensorSpec containing the expected inputs. Can be `None`, in which case the signature will be inferred from the model. Raises: ValueError: If the train/eval mode is being exported, but the model does not have an optimizer. zModel z0 does not have an optimizer. Cannot export mode .N) input_tensorsr! compile_clonerTr5) allow_empty)signature_def_mapsaverinit_optrain_op)0rrJrMrFrCrXrr.rget_default_graphGraph as_defaultrlearning_phase_scoperKnest map_structurer[r clone_and_build_modeladd_to_collection GraphKeys GLOBAL_STEP iterations_make_train_functiontrain_function updates_oprL_make_test_function_make_predict_functionget_collection_ref UPDATE_OPSextend state_updatesSessionrU load_weightsassert_existing_objects_matched"_assert_same_non_optimizer_objectsr:_has_saved_variablesadd_meta_graph model_utilsEXPORT_TAG_MAP_create_signature_def_maptrainSaverlocal_variables_initializer)moderOr=rr!r>r#r^ model_graphgr]clonercZclone_var_liststatusr%r%r&rIst               & rIc Cs&ddt|j|jD}|jr8dd|jD}||ddt|j|jD}t |}t t j j t j j jj}t }|dur|D]6\}} t| tjr|| j| | jdf||<q||}|D]} t j j t j j jj| qtj|||jr|jnd|d} tj|| |tjj kdS) z.Creates a SignatureDef map from a Keras model.cSsi|]\}}||qSr%r%.0rXxr%r%r& pz-_create_signature_def_map..cSs&i|]}|dur|jdd|qS)N:r)rXsplit)rrr%r%r&rrscSsi|]\}}||qSr%r%rr%r%r&rvsNr) predictionslossmetrics)export_outputsr$)!zip input_namesinputsrF_targetsupdate output_namesoutputsrextract_model_metricssetrr.rget_collectionrlLOCAL_VARIABLESitemsr@r Metric variablesresultupdates differencerkr~export_outputs_for_mode total_lossbuild_all_signature_defsrrJrM) rrZ inputs_dict targets_dictZ outputs_dictr local_varsZ vars_to_addkeyvaluevrr%r%r&rnsF        rcCsdS)z;Asserts model and clone contain the same trackable objects.Tr%)rrrZ clone_graphr%r%r&r{sr{z(keras.experimental.load_from_saved_modelcCstjdddtjjtj|tjtjj tjt }tjj |d}| }Wdn1sj0Yt j||d}tjjtj|tjtjjtjtjj}|||S)a,Loads a keras Model from a SavedModel created by `export_saved_model()`. This function reinstantiates model state by: 1) loading model topology from json (this will eventually come from metagraph). 2) loading model weights from checkpoint. Example: ```python import tensorflow as tf # Create a tf.keras model. model = tf.keras.Sequential() model.add(tf.keras.layers.Dense(1, input_shape=[10])) model.summary() # Save the tf.keras model in the SavedModel format. path = '/tmp/simple_keras_model' tf.compat.v1.keras.experimental.export_saved_model(model, path) # Load the saved keras model back. new_model = tf.compat.v1.keras.experimental.load_from_saved_model(path) new_model.summary() ``` Args: saved_model_path: a string specifying the path to an existing SavedModel. custom_objects: Optional dictionary mapping names (strings) to custom classes or functions to be considered during deserialization. Returns: a keras.Model instance. z`tf.keras.experimental.load_from_saved_model` is deprecatedand will be removed in a future version. Please switch to `tf.keras.models.load_model`.rrrN)r!)rrrr*r+r,r.as_bytesrASSETS_DIRECTORYr/r0readrmodel_from_jsonr"VARIABLES_DIRECTORYVARIABLES_FILENAMEry)r r!r3r4r2rr;r%r%r&load_from_saved_models*%  &  rcCst|}tjj||S)zBReturn variables sub-directory, or create one if it doesn't exist.)_get_variables_dirrr*r+makedirs) export_dir variables_dirr%r%r&r8sr8cCs$tjjtj|tjtjjS)z1Return variables sub-directory in the SavedModel.)rr*r+r,r.r"rrrr%r%r&rs rcCs(tjjtjt|tjtjjS)zCReturn the variables path, used as the prefix for checkpoint files.) rr*r+r,r.r"rrrrr%r%r&r9sr9cCst|}tjj||S)z?Return assets sub-directory, or create one if it doesn't exist.)_get_assets_dirrr*r+r)rassets_destination_dirr%r%r&r-sr-cCs$tjjtj|tjtjjS)z1Return path to asset directory in the SavedModel.)rr*r+r,r.r"rrrr%r%r&rs r)NFNF)N).__doc__rZtensorflow.compat.v2r.v2rkerasrZkeras.optimizersrkeras.optimizers.optimizer_v2rZ keras.savingrrrr~ keras.utilsrkeras.utils.generic_utilsr tensorflow.python.platformr r tensorflow.python.util.tf_exportr globalsr r rr/r'rr<rrUr[rIrr{rr8rr9r-rr%r%r%r&sJ            ` Dw.  C