ёi5SSKJr SSKrSSKrSSKrSSKrSSKrSSKrSSKrSSK J r J r J r J r SSKrSSKrSSKJrJrJrJrJr SSKJrJr SSKJrJrJrJrJr SSK J!r! SSK"J#r#J$r$J%r%J&r&J'r'J(r(J)r)J*r* S S K+J,r,J-r-J.r.J/r/J0r0 S S K1J2r2J3r3J4r4J5r5J6r6J7r7J8r8J9r9 \ (a^SS K:J;r;J SS K?J@r@JArA SSKJBrB "SS\ 5rC"SS\ 5rD"SS\ 5rE"SS\ 5rF"SS\ 5rG"SS\ 5rH/rI\!"\J\RSS9rLSrMS4S5SjjrNS6S7SjjrOS8S jrP\S9S!j5rQS:S;S"jjrR\SS%jrU\S?S&j5rV\S@S'j5rW\SAS(j5rXSBS)jrY\SCS*j5rZ\ \SDSES+jj55r[\ \SDSFS,jj55r[\SGS-j5r[SGSHS.jjr\\SISJS/jj5r]\SKSLS0jj5r^\SMS1j5r_\SNS2j5r`SOSPS3jjrag)Q) annotationsN) TYPE_CHECKINGAny TypedDictoverload)ProgramVariablecoredefault_main_program unique_name)Executor global_scope) Parameterdygraph_not_support in_pir_modeprocess_type_promotion static_only) get_logger)_clone_var_in_block__load_program_scope_pack_loaded_dict_pickle_loads_mac_unpack_saved_dictis_belong_to_optimizer is_parameteris_persistable) _check_args _check_vars_get_valid_program_normalize_path_prefix_safe_load_pickle)get_pir_parametersload_inference_model_pirload_pir load_vars_pirnormalize_pir_programsave_inference_model_pirsave_pir save_vars_pir)CallableSequence) NotRequiredUnpack)Tensorc \rSrSr%S\S'Srg)_NormalizeProgramKwargsQNotRequired[bool]skip_prune_programN__name__ __module__ __qualname____firstlineno____annotations____static_attributes__r5P/var/www/html/banglarbhumi/venv/lib/python3.13/site-packages/paddle/static/io.pyr1r1Qs--r=r1c*\rSrSr%S\S'S\S'Srg)_SerializeProgramKwargsTNotRequired[Program]programr3 legacy_formatr5Nr6r5r=r>r@r@Ts%%((r=r@c \rSrSr%S\S'Srg)_SerializePersistablesKwargsXrBrCr5Nr6r5r=r>rFrFXs%%r=rFc4\rSrSr%S\S'S\S'S\S'Srg) _SaveInferenceModelKwargs[rBrCr3 clip_extrarDr5Nr6r5r=r>rIrI[s%%%%((r=rIc*\rSrSr%S\S'S\S'Srg)_LoadInferenceModelKwargs`zNotRequired[str]model_filenameparams_filenamer5Nr6r5r=r>rMrM`s(())r=rMc \rSrSr%S\S'Srg) _SaveKwargsdzNotRequired[int]pickle_protocolr5Nr6r5r=r>rRrRds))r=rRz&%(asctime)s-%(levelname)s: %(message)s)fmtc [U[5(deURR5[R R R:XaFURURURURURURSS9$URURURURURSS9$)NTnameshapedtypetype lod_level persistable)rXrYrZr[r]) isinstancer descr[r VarDescVarType DENSE_TENSOR create_varrXrYrZr\blockvars r>_clone_var_in_blockrgos c8 $ $$ $ xx}}$,,..;;;))))mm    ))))    r=c [U5S:XagUR5nURU[RR R SS9n[U5H^upVURU5(d[SUSUSUSUS3 5eURU5nURS S U/0S U/0S U0S 9 M` g)NrTrXr[r]zThe feeded_var_names[z]: 'zC' doesn't exist in pruned inference program. Please check whether 'zC' is a valid feed_var name, or remove it from feeded_var_names if 'z1' is not involved in the target_vars calculation.feedXOutcolr[inputsoutputsattrs) len global_blockrcr r`raFEED_MINIBATCH enumeratehas_var ValueErrorrf _prepend_op)inference_programfeed_target_namesfeed_holder_namersfeed_varirXouts r>prepend_feed_opsrs  "$113L&&  \\ ! ! 0 0'H ./##D))'s$tf5))-/fMO  t$  ($SEN!* ! 0r=cUR5nURU[RRR SS9n[ U5HupVURSSU/0SU/0SU0S9 M! g)NTrifetchrkrlrmrn)rsrcr r`ra FETCH_LISTru append_op)ryfetch_target_namesfetch_holder_namers fetch_varr}rXs r>append_fetch_opsrs %113L''  \\ ! ! , ,(I /0$=YK(!*  1r=c [5(a [XU40UD6$[U[5(d[ S[ U5S35e[U[ 5(dU/n[SU55(d [ S5e[U[ 5(dU/n[SU55(d [ S5e[UR5R5S:Xa [S5eUR5RH[n[RR5nURUS 5 UR S :XdME[ R""S 5 O UR%5nUR5n/n['UR5HupUR(R+S 5 UR S :XdUR S:XaUR-U 5 UR S:XdMcUR/S5n [U 5S:dMU Sn UR0R3U 5 /n U SSH#n U R-UR5U 55 M% UR7SU 5 M USSS2HnUR9U5 M UR(R;5 UVs/sHoR<PM nnUR?SS 5nU(dURAUUS9nURCSS9nUVs/sHoR<PM nn[EUU5 [GUU5 UR(RI5 U$s snfs snf)a- Normalize/Optimize a program according to feed_vars and fetch_vars. Args: program(Program): Specify a program you want to optimize. feed_vars(Tensor | list[Tensor]): Variables needed by inference. fetch_vars(Tensor | list[Tensor]): Variables returned by inference. kwargs: Supported keys including ``skip_prune_program``. - skip_prune_program(bool): whether to skip pruning program. Defaults to False. Returns: Program: Normalized/Optimized program. Examples: .. code-block:: python >>> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # User defined network, here a softmax regression example >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # normalize main program. >>> program = paddle.static.default_main_program() >>> normalized_program = paddle.static.normalize_program(program, [image], [predict]) 3program type must be `base.Program`, but received ``c3B# UHn[U[5v M g7fNr^r .0vs r> $normalize_program..s: 1z!X&& z8feed_vars type must be a Variable or a list of Variable.c3B# UHn[U[5v M g7frrrs r>rrs; 1z!X&& rz9fetch_vars type must be a Variable or a list of Variable.rz=program must not be empty. at least one operator is required!auczHBe sure that you have set auc states to 0 before saving inference model.FrjrpylayerblocksrNr4)feeded_var_namestargetsT) prune_read_op)%rr'r^r TypeErrorr[listallrrrsopsrwr op_proto_and_checker_makerkOpDeviceAttrName _set_attrwarningswarnclonerur_ set_is_targetappend_blocks_attr_idsrpopre_update_desc_attr _remove_opflushrXget_prune_with_input_inference_optimizerr _set_version)rC feed_vars fetch_varskwargsopdevice_attr_name copy_programrs remove_op_idxr}sub_blocks_idsbackward_block_idreserved_blocksblock_ididxrffeed_var_namesr4fetch_var_namess r>normalize_programrsX}}$WNvNN gw ' 'A$w-PQ R   i & &K : : : : F   j$ ' ' \ ; ; ; ; G   7   ! % %&!+ K  ""$((::LLN %r* 77e  MMZ  )==?L,,.LM<++, e$ 77f 7 2   # 77i 00:N>"Q&$22$6!##''(9:"$ .s 3H#**<+=+=h+GH!4$$X?!-$TrT"$#*34)3hh)N4$8%@ #55+Z6  33$3GL+56:Cxx:O6\>2\?3""$ 57s M%M*c [SU5 [SU5 [URSS55n[X0U5nURSS5n[ X4S9$)a Serialize default main program according to feed_vars and fetch_vars. Args: feed_vars(Tensor | list[Tensor]): Tensor needed by inference. fetch_vars(Tensor | list[Tensor]): Tensor returned by inference. kwargs: Supported keys including ``program``. Attention please, kwargs is used for backward compatibility mainly. - program(Program): specify a program if you don't want to use default main program. - legacy_format(bool): whether to save inference program in legacy format. Defaults to False. Returns: bytes: serialized program. Examples: .. code-block:: python >>> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # User defined network, here a softmax regression example >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # serialize the default main program to bytes. >>> serialized_program = paddle.static.serialize_program([image], [predict]) # deserialize bytes to program >>> deserialized_program = paddle.static.deserialize_program(serialized_program) rrrCNrDFrD)rr rr_serialize_program)rrrrCrDs r>serialize_programr0sU^ Y' j) It!<=GJ?GJJ6M g CCr=c4URRUS9$)z# serialize given program to bytes. r)r_serialize_to_string)rCrDs r>rris << + +- + HHr=c [SU5 [SU5 [URSS55n[X@U5n[ XB5$)a Serialize parameters using given executor and default main program according to feed_vars and fetch_vars. Args: feed_vars(Tensor | list[Tensor]): Tensor needed by inference. fetch_vars(Tensor | list[Tensor]): Tensor returned by inference. kwargs: Supported keys including ``program``. Attention please, kwargs is used for backward compatibility mainly. - program(Program): specify a program if you don't want to use default main program. Returns: bytes: serialized program. Examples: .. code-block:: python >>> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # User defined network, here a softmax regression example >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # serialize parameters to bytes. >>> serialized_params = paddle.static.serialize_persistables([image], [predict], exe) # deserialize bytes to parameters. >>> main_program = paddle.static.default_main_program() >>> deserialized_params = paddle.static.deserialize_persistables(main_program, serialized_params, exe) rrrCN)rr rr_serialize_persistables)rrexecutorrrCs r>serialize_persistablesrpsE` Y' j) It!<=GJ?G "7 55r=cr[[[UR555n[ U5S:Xa[ R "S5 g[5nUR5n0nUHPnUR[RRR:wdM7[XF5nXeUR'MR /n[!UR#55Hn UR%XY5 M [&R("S5n UR+[RRRU S9n U R,R/S5 UR1SSU0S U 0S SS .S 9 UR35 UR5U5 [75R9U 5R;5$) z8 Serialize parameters using given program and executor. rVno variable in your model, please ensure there are any variables in your model to saveNout_varr[rXT save_combinerkYr file_pathsave_to_memoryrn)rfilterr list_varsrrrrrrsr[r r`raRAWrgrXsortedkeysrr generatercr_set_persistabler_sync_with_cpprunrfind_var get_bytes) rCrvars_ save_program save_block save_var_maprfvar_copyin_varsrX out_var_namers r>rrsu (9(9(;< =E 5zQ  . 9L**,JL 88t||++// /*:;H*- ' G|((*+|)*,'' 2L## \\ ! ! % %L$G LL  & W~g$7 ! LL > " "< 0 : : <>> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # 用户自定义网络，此处用 softmax 回归为例。 >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # 序列化参数 >>> serialized_params = paddle.static.serialize_persistables([image], [predict], exe) # 将序列化之后的参数保存到文件 >>> params_path = path_prefix + ".params" >>> paddle.static.save_to_file(params_path, serialized_params) z'content' type should be bytes.wbN)r^bytesrwopenwrite)pathcontentfs r> save_to_filersBD gu % %:;; dD Q    s A Ac [5(a[XX#40UD6 g[U5n[RR U5n[R "U5 US-nUS-n[RRU5(a[SUS35e[RRU5(a[SUS35e[SU5 [SU5 [URSS55n [U 5n URS S 5n [!U UUURS S 5S 9n URSS 5n [#U R%U S9U S9n ['URSS 5(a=[RR)[RR U5S5OUU 5 [+[-[.U R1555n [3[+U 55S:Xa[4R6"S5 [3U 5S:ai[RR U5n[RR9U5n[;UUU [.URSS 5(aSOUS9 gg![a*nUR[R:waeSnAGNbSnAff=f)aa Save current model and its parameters to given path. i.e. Given ``path_prefix = "PATH/modelname"``, after invoking ``save_inference_model(path_prefix, feed_vars, fetch_vars, executor)``, you will find two files named ``modelname.pdmodel`` and ``modelname.pdiparams`` under ``PATH``, which represent your model and parameters respectively. Args: path_prefix(str): Directory path to save model + model name without suffix. feed_vars(Tensor | list[Tensor]): Variables needed by inference. fetch_vars(Tensor | list[Tensor]): Variables returned by inference. executor(Executor): The executor that saves the inference model. You can refer to :ref:`api_guide_executor_en` for more details. kwargs: Supported keys including 'program' and "clip_extra". Attention please, kwargs is used for backward compatibility mainly. - program(Program): specify a program if you don't want to use default main program. - clip_extra(bool): the flag indicating whether to clip extra information for every operator. Default: True. - legacy_format(bool): whether to save inference model in legacy format. Default: False. Returns: None Examples: .. code-block:: python >>> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # User defined network, here a softmax regression example >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # Feed data and train process # Save inference model. Note we don't save label and loss in this example >>> paddle.static.save_inference_model(path_prefix, [image], [predict], exe) # In this example, the save_inference_mode inference will prune the default # main program according to the network's input node (img) and output node(predict). # The pruned inference program is going to be saved in file "./infer_model.pdmodel" # and parameters are going to be saved in file "./infer_model.pdiparams". N.pdmodel .pdiparams'z' is an existing directory.rrrCrKTr4F)r4rD)rKrseparate_parameters __model__rrdirname main_program predicatefilename)rr(r!osrrmakedirsOSErrorerrnoEEXISTisdirrwrr rrrr_remove_training_inforjoinrrrrrrrrbasename save_vars) path_prefixrrrrre model_path params_pathrCrKrD program_bytesvars save_dirnamerPs r>save_inference_modelr si~}} J >> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # User defined network, here a softmax regression example >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # serialize the default main program to bytes. >>> serialized_program = paddle.static.serialize_program([image], [predict]) # deserialize bytes to program >>> deserialized_program = paddle.static.deserialize_program(serialized_program) zUnsupported program version:  )rparse_from_stringr _is_program_version_supported_versionrw)datarCs r>deserialize_programrsSL''-G  - -g.>.>.@ A A89I9I9K8LBOPP Nr=c [U[5(d[S[U5S35e[5nUR 5n[ [ [UR555n0n0n/n/n UGHn [U [5(deU R[RRR:XaMO[U [5(a/[U R R#55XjR$'U R[RRR&:XaU R)U 5 M[+XJ5n UR)U 5 XU R$'GM Uc[-U5S:XdS5eg/n [/UR155Hn U R)X}5 M UR3S0SU 0USS .S 9 UR5U5 UHn [U [5(dM[6R8R;5R=U R$5nUcS U R$-5e[>R@"URC55RDnU R$U;dU R$S -5eURGU R$5nUU:wdM[IS USU R$SUS35e g)a# Deserialize given data to parameters according to given program and executor. Args: program(Program): program that contains parameter names (to deserialize). data(bytes): serialized parameters. executor(Executor): executor used to run load op. Returns: Program: deserialized program. Examples: .. code-block:: python >>> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # User defined network, here a softmax regression example >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # serialize parameters to bytes. >>> serialized_params = paddle.static.serialize_persistables([image], [predict], exe) # deserialize bytes to parameters. >>> main_program = paddle.static.default_main_program() >>> deserialized_params = paddle.static.deserialize_persistables(main_program, serialized_params, exe) rrNrz]Required 'data' shall be not None if program contains parameter, but received 'data' is None. load_combinerlTrmodel_from_memoryrncan't not find var: z MUST in var list.z6Shape mismatch, program needs a parameter with shape (z), but the loaded parameter ('z') has a shape of ().)%r^rrr[rsrrrrr r r`rarrtupler_ get_shaperX SELECTED_ROWSrrgrrrrrrpaddlebaserrnparray get_tensorrYr RuntimeError)rCrr load_program load_blockrorigin_shape_map load_var_map check_vars sparse_varsrfr load_var_listrXvar_tmp new_shape origin_shapes r>deserialize_persistablesr,sX gw ' 'A$w-PQ R  9L**,J (9(9(;< =ELJK#x(((( 88t||++// /  c9 % %).sxx/A/A/C)D XX & 88t||++99 9   s # &z7#&.X]]# |#$) k ) M|((*+\/0,  & t<  LL#y)) ++**,55chh?"E$:SXX$EE"XXg0023:: xx++LSXX8L-LL+'++CHH5  $HW..1hhZ7J9+UWY r=ct[US5nUR5nSSS5 U$!,(df  W$=f)u Load file in binary mode. Args: path(str): Path of an existed file. Returns: bytes: Content of file. Examples: .. code-block:: python >>> import paddle >>> paddle.enable_static() >>> path_prefix = "./infer_model" # 用户自定义网络，此处用 softmax 回归为例。 >>> image = paddle.static.data(name='img', shape=[None, 28, 28], dtype='float32') >>> label = paddle.static.data(name='label', shape=[None, 1], dtype='int64') >>> predict = paddle.static.nn.fc(image, 10, activation='softmax') >>> loss = paddle.nn.functional.cross_entropy(predict, label) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(paddle.static.default_startup_program()) # 序列化参数 >>> serialized_params = paddle.static.serialize_persistables([image], [predict], exe) # 将序列化之后的参数保存到文件 >>> params_path = path_prefix + ".params" >>> paddle.static.save_to_file(params_path, serialized_params) # 从文件加载序列化之后的参数 >>> serialized_params_copy = paddle.static.load_from_file(params_path) rbN)rread)rrrs r>load_from_filer0$s4H dD Qvvx  K   Ks( 7c X [5(a [X40UD6$SnSn[R"5RR n[ XRX45 Uc[RS5 URSS5nURSS5nUc [S5eUn[U5n [U 5n [[[U R!555n [#U 5S:a[%USU [US 9 GOf['U5n[(R*R-U5n [(R*R/U 5(d[S U 35eU(d US -n US -n GOQURSS5nURSS5nUc![(R*R1US 5n Oe[(R*R1XS -5n [(R*R3U 5(d[(R*R1X5n Uc![(R*R1US5n Oe[(R*R1XS -5n [(R*R3U 5(d[(R*R1X5n [RSU SU 35 [5U 5n[U5n [U 5n [[[U R!555n [#U 5S:aO[(R*R-U 5n[(R*R7U 5n[%UUU [US 9 U R8R;5n[<R>RA5(Ga[<RBRE5 [<RFRIU R85n U RK5n/n/nURLHnURO5S:XdMURQ5SnURSS5nU [<RTRWUURXURZS9nUR]U5 UR_5RaU5 SSS5 URcU5 M UHnUReU5 M URLH9nURO5S:XdMURcURgS55 M; SSS5 OIU R8Ri5nUVs/sH"nU RK5RkU5PM$ nnXW/$!,(df  N=f!,(df  N#=fs snf)a Load inference model from a given path. By this API, you can get the model structure(Inference Program) and model parameters. Args: path_prefix(str | None): One of the following: - Directory path to save model + model name without suffix. - Set to None when reading the model from memory. executor(Executor): The executor to run for loading inference model. See :ref:`api_guide_executor_en` for more details about it. kwargs: Supported keys including 'model_filename', 'params_filename'. Attention please, kwargs is used for backward compatibility mainly. - model_filename(str): specify model_filename if you don't want to use default name. - params_filename(str): specify params_filename if you don't want to use default name. Returns: list: The return of this API is a list with three elements: (program, feed_target_names, fetch_targets). The `program` is a ``Program`` (refer to :ref:`api_guide_Program_en`), which is used for inference. The `feed_target_names` is a list of ``str``, which contains names of variables that need to feed data in the inference program. The `fetch_targets` is a list of ``Variable`` (refer to :ref:`api_guide_Program_en`). It contains variables from which we can get inference results. Examples: .. code-block:: python >>> import paddle >>> import numpy as np >>> paddle.enable_static() # Build the model >>> startup_prog = paddle.static.default_startup_program() >>> main_prog = paddle.static.default_main_program() >>> with paddle.static.program_guard(main_prog, startup_prog): ... image = paddle.static.data(name="img", shape=[64, 784]) ... w = paddle.create_parameter(shape=[784, 200], dtype='float32') ... b = paddle.create_parameter(shape=[200], dtype='float32') ... hidden_w = paddle.matmul(x=image, y=w) ... hidden_b = paddle.add(hidden_w, b) >>> exe = paddle.static.Executor(paddle.CPUPlace()) >>> exe.run(startup_prog) # Save the inference model >>> path_prefix = "./infer_model" >>> paddle.static.save_inference_model(path_prefix, [image], [hidden_b], exe) >>> [inference_program, feed_target_names, fetch_targets] = ( ... paddle.static.load_inference_model(path_prefix, exe)) >>> tensor_img = np.array(np.random.random((64, 784)), dtype=np.float32) >>> results = exe.run(inference_program, ... feed={feed_target_names[0]: tensor_img}, ... fetch_list=fetch_targets) # In this example, the inference program was saved in file # "./infer_model.pdmodel" and parameters were saved in file # " ./infer_model.pdiparams". # By the inference program, feed_target_names and # fetch_targets, we can use an executor to run the inference # program to get the inference result. )rOrP)pserver_endpointsNzKLoad inference model from memory is deprecated. Please specify path_prefix.rOrPz8params_filename cannot be None when path_prefix is None.rrzThere is no directory named rrrrz[The old way to load inference model is deprecated. Please specify path_prefix. model path: z, params path: z pd_op.feedrX)rXrYrZz pd_op.fetch)6rr$inspect currentframef_codeco_namer_loggerwarningrrwrrrrrrrr load_varsr!rrrrrexistsr0rr_get_feed_target_namesr frameworkin_pir_executor_mode pir_utilsIrGuardpirtranslate_to_pirrsrrXrqresultstaticrrYrZreplace_all_uses_withget_defining_op move_beforer remove_opoperand_sourceget_fetch_target_namesrf)rrrsupported_argsdeprecated_argscallerrOrPrrCr dir_pathrr load_dirnamerzreremove_op_list fetch_targetsrvar_name org_valuevaluerrXs r>load_inference_modelrTMsL}}' HHH:N,O  ! ! # * * 2 2F@ Y  $4d; **%6=  "J ' %m4)1F>7+<+<+>?@ t9q= $((  -[9 77??;/ww}}X&&;H:FG G$z1J% 4K$ZZ(8$?N$jj):DAO%WW\\+{C WW\\*!< ww~~j11!#k!JJ& ggll;;  ggll7+<+<+>?@ t9q=77??;7L gg..{;O $$((   ::< ,,..    % % 'jj11',,?G((*ENMii779 ,!xxz&1H " ! I & 2 2!)"+//"+//!3! "77>--/;;B?#))"- %#%ii779 -!(():):1)=> )( '2%\\@@B9K 9KG " & &t ,9K    66)( '4 s9AV%&V A$V/AV $V)V' V V V$cgrr5rrrr rrs r>rrs r=cgrr5rVs r>rr!s r=c [5(a [XX55$SnUcUcSn[U5nUc-[UUU[ [ XBR 555US9$Sn[[ U55S:Xa[R"S5 g[5nUR5n 0n UHn U R[RRR :XaM7[#X5n UcUSLa}[$R&R)[$R&R+U5U R,5n U R/SS U /00S [$R&R+U 50S 9 MXU R,'M UcU(a/n[1U R355HnUR5X5 M S nUSLa=[$R&R)[$R&R+U5U5nU R7[RRR US 9nUR8R;S5 U R/SS U0SU0UUS.S 9 UR=5 UR?5 URAU5 U(a'[C5REU5RG5$g)a Save specific variables in the `Program` to files. There are two ways to specify the variables to be saved: set variables in a list and assign it to the `vars`, or use the `predicate` function to select variables that make `predicate(variable) == True`. The first way has a higher priority. The `dirname` is used to specify the folder where to save variables. If you prefer to save variables in separate files in the `dirname` folder, do not set `filename`. If you prefer to save all variables in a single file, use `filename` to specify it. Args: executor(Executor): The executor to run for saving variables. dirname(str, optional): The folder where to save variables. When you need to save the parameter to the memory, set it to None. main_program(Program, optional): The program whose variables will be saved. If it is None, the default main program will be used automatically. Default: None vars(list[Variable], optional): The list contains all variables to be saved. Default: None predicate(function, optional): The function selects the variables that make `predicate(variable) == True`. Default: None filename(str, optional): If you prefer to save all variables in a single file, use `filename` to specify it. Otherwise, let `filename` be None. Default: None Returns: str: When saving parameters to a file, returns None. When saving parameters to memory, returns a binary string containing parameters. Raises: TypeError: If `main_program` is not an instance of Program nor None. Examples: .. code-block:: python >>> import paddle >>> import paddle.static as static >>> paddle.enable_static() >>> main_prog = static.Program() >>> startup_prog = static.Program() >>> with static.program_guard(main_prog, startup_prog): ... data = paddle.static.data(name="img", shape=[64, 784]) ... w = paddle.create_parameter(shape=[784, 200], dtype='float32', name='fc_w') ... b = paddle.create_parameter(shape=[200], dtype='float32', name='fc_b') ... hidden_w = paddle.matmul(x=data, y=w) ... hidden_b = paddle.add(hidden_w, b) >>> place = static.CPUPlace() >>> exe = static.Executor(place) >>> exe.run(startup_prog) # The first usage: use `vars` to set the saved variables. >>> var_list = [w, b] >>> path = "./my_paddle_vars" # w and b will be save in a file named "var_file". >>> paddle.static.io.save_vars(executor=exe, dirname=path, vars=var_list, ... filename="vars_file") # The second usage: use `predicate` to select the saved variable. >>> def name_has_fc(var): ... res = "fc" in var.name ... return res >>> param_path = "./my_paddle_model" # all variables whose names contain "fc " are saved. >>> paddle.static.io.save_vars(executor=exe, dirname=param_path, main_program=main_prog, vars=None, predicate = name_has_fc) FNT)rrr r saved_paramsrrsaverkrrnrrrrr)$rr*r rrrrrrrrrrsr[r r`rarrrrrnormpathrXrrrrrcr_rrrrrrr)rrrr rrrparams_var_namerrreach_varnew_varsave_file_path save_var_listrX save_pathrYs r>rr-sf}}WDCCN8+%l3L | %fY(>(>(@AB   ) tDz?a  MMh y !..0  H}} 4 4 8 88*:@GNe$;!#GG$$W-w||"$$'+&(8(8(HI %.5W\\*!$  >M|0023$$\%784I&GGLL)9)9')BHM %00\\))--O1L    - -d 3  #]+l+!*&4 !  ##% \" >**?;EEG G r=c  [5(a [XX#U5$SnUb [RR U5nOSnUS:XaSnUcgUc [ 5n[ U[5(d[S[U535e[UUU[[XBR555US9 g[5nUR5nUc [ 5n[ U[5(d[S[U535e0n 0n /n /n UGHTn [ U [5(deU R[ R"R$R&:XaMO[ U [(5(a/[+U R,R/55XR0'U R[ R"R$R2:XaU R5U 5 M[7X5nU R5U 5 UcOUc [9S5eUR;S0S U/0S [RR=XR050S 9 GMFXUR0'GMW U GHn [ U [5(deUb [9S 5e[7X5n[RR=XR05n[RR?U5(d UR0SU35e[RRAU5(aAUR;S0S U/0S [RR=XR050S 9 GM/n[RB"U5nUH6nUREUR05(dM%UR5U5 M8 /nUH}nURGUURURHURJSS9nUR5U5 [RR=UUS5nUR;S0S U/0S U0S 9 M UR;SSU0S U00S 9 GM Ubq/n[MU RO55HnUR5U U5 M USLa[RR=X5nUR;S0S U0UUS.S 9 URQU5 U Hn [ U [(5(dM[RRTRW5RYU R05nUcSU R0-5e[ZR\"UR_55RHnU R0U ;dU R0S-5eU RaU R05nUU:wdM[cSUSU R0SUS35e g)a :api_attr: Static Graph This API loads variables from files by executor. There are two ways to specify the variables to be loaded: the first way, set variables in a list and assign it to the `vars`; the second way, use the `predicate` function to select variables that make `predicate(variable) == True`. The first way has a higher priority. The `dirname` is used to specify the folder where to load variables. If variables were saved in separate files in the folder `dirname`, set `filename` None. If all variables were saved in a single file, use `filename` to specify it. Args: executor(Executor): The executor to run for loading variables. dirname(str): The folder where to load the variables. main_program(Program, optional): The program whose variables will be loaded. If it is None, the default main program will be used automatically. Default: None vars(list[Variable], optional): The list that contains all variables to be loaded. Default: None predicate(function, optional): The function selects variables that make `predicate(variable) == True`. Default: None filename(str, optional): The file which saved all required variables. If variables were saved in separate files, set it to be None. Default: None Returns: None Examples: .. code-block:: python >>> import paddle >>> import paddle.static as static >>> paddle.enable_static() >>> main_prog = static.Program() >>> startup_prog = static.Program() >>> with static.program_guard(main_prog, startup_prog): ... data = paddle.static.data(name="img", shape=[64, 784]) ... w = paddle.create_parameter(shape=[784, 200], dtype='float32', name='fc_w') ... b = paddle.create_parameter(shape=[200], dtype='float32', name='fc_b') ... hidden_w = paddle.matmul(x=data, y=w) ... hidden_b = paddle.add(hidden_w, b) >>> place = paddle.CPUPlace() >>> exe = static.Executor(place) >>> exe.run(startup_prog) # The first usage: using `vars` to specify the variables. >>> path = "./my_paddle_vars" >>> var_list = [w, b] >>> paddle.static.io.save_vars(executor=exe, dirname=path, vars=var_list, ... filename="vars_file") >>> paddle.static.io.load_vars(executor=exe, dirname=path, vars=var_list, ... filename="vars_file") # w and b will be loaded, and they are supposed to # be saved in the same file named 'var_file' in the path "./my_paddle_vars". # The second usage: using the `predicate` function to select variables >>> param_path = "./my_paddle_model" >>> def name_has_fc(var): ... res = "fc" in var.name ... return res >>> paddle.static.io.save_vars(executor=exe, dirname=param_path, main_program=main_prog, ... vars=None, predicate=name_has_fc) >>> paddle.static.io.load_vars(executor=exe, dirname=param_path, main_program=main_prog, ... vars=None, predicate=name_has_fc) # Load All variables in the `main_program` whose name includes "fc". # And all the variables are supposed to be saved in separate files. FNTrzWThe type of input main_program is invalid, expected type is base.Program, but received )rrr rz>The directory path and params cannot be None at the same time.loadrlrrnz.SelectedRows can not be load with load_combinezSelectedRows var z can not find at )rXr[rYrZr]Paramlookup_sparse_table_mergerkrrrzMUST in var listzUVariable's shape does not match, the Program requires a parameter with the shape of ((), while the loaded parameter (namely [  ]) has a shape of (r)2rr&rrr[r r^rrr[r9rrrrsr r r`rarrrr_rrXrrrrwrrr:isfilelistdir startswithrcrYrZrrrrrrrrrr rr!)rrrr rrvars_from_memory load_progr#orig_para_shaper%r&r'r]r^var_pathr block_pathsreslicesslicerr(rXvar_tempr* orig_shapes r>r9r9sl}}X HMM''""7+2~ |  /1L,00ijno{j|i}~   %fY(>(>(@AB  I ++-  /1L,00ijno{j|i}~     Hh11 11}} 4 4 8 88(I..16MM++-2 .}} 4 4 B BB""8,*:@G   h '?$X$$"WI.& Wll(KL %.5W\\*=@$Hh11 11# D+:@Gww||G\\:H77>>(++ ' ~5FxjQww~~h''$$"WI.& Wll(KL % jj2 (E'' 55 e,)#E&11"$\\%mm%mm$) 2EMM%( " Xug FI((#!!& 0*I6 )$$$$4="G, %e$r  M|0023$$\$%784 5(77<<:  # .!))9 !   Y#Hh 22{{//1::8==IH' O)?(--)O O'("5"5"78??I==O3  22 3),,X]];JJ&"klvkwx<UR#555n U Vs0sHoR$U"U5_M n n[5US-S5n [0R<"XUS 9 SSS5 URA5n URBRE5 [5US-S5n U RGURBRI55 SSS5 gs snf!,(df  N=f!,(df  GN=fs snf!,(df  N=f!,(df  g=f)a This function save parameters, optimizer information and network description to model_path. The parameters contains all the trainable Tensor, will save to a file with suffix ".pdparams". The optimizer information contains all the Tensor used by optimizer. For Adam optimizer, contains beta1, beta2, momentum etc. All the information will save to a file with suffix ".pdopt". (If the optimizer have no Tensor need to save (like SGD), the fill will not generated). The network description is the description of the program. It's only used for deployment. The description will save to a file with a suffix ".pdmodel". Args: program(Program) : The program to saved. model_path(str): the file prefix to save the program. The format is "dirname/file_prefix". If file_prefix is empty str. A exception will be raised protocol(int, optional): The protocol version of pickle module must be greater than 1 and less than 5. Default: 4 configs(dict, optional) : optional keyword arguments. Returns: None Examples: .. code-block:: python >>> import paddle >>> import paddle.static as static >>> paddle.enable_static() >>> x = static.data(name="x", shape=[10, 10], dtype='float32') >>> y = static.nn.fc(x, 10) >>> z = static.nn.fc(y, 10) >>> place = paddle.CPUPlace() >>> exe = static.Executor(place) >>> exe.run(static.default_startup_program()) >>> prog = static.default_main_program() >>> static.save(prog, "./temp") rzThe input model_path MUST be format of dirname/filename [dirname\filename in Windows system], but received model_path is empty string.rTzJ'pickle_protocol' is a deprecated argument. Please use 'protocol' instead.z+The 'protocol' MUST be `int`, but received z/Expected 1<'protocol'<5, but received protocol=c[5RUR5R5n[R "U5$r)rrrXr rr)rfts r>r save..get_tensors/ N # #CHH - 8 8 :xx{r=darwin)protocol .pdparamsri@c32># UH nTXT-v M g7frr5)rr} max_bytes pickle_bytess r>rsave..)s!?AQY/?srN.pdoptr)%rr)rrrrrr^intrwr[rr:rrrrrrXrsysplatform version_infomajorpickledumpsr writelinesrangerrdumprrr_rrr)rCrr|configs base_namedir_namer parameter_listp param_dictroptimizer_var_listopt_dictrrrs @@r>rZrZsX}}XAAA  ,I ? R?G#,- X  h $ $9$x.9I J  !|x!|=hZ H  wwz*Hx00 H&w/@/@/BCDN1?@A&&*Q-'J@#J9J ||xC$4$4$:$:a$?||JB *{*D 1QI LLq#l"3Y? 2 1*{*D 1Q KK  92%w'8'8':;0BB/A! 1 %/AHB j8#T *a H(3 +==?L LL j:%t ,  0023 - ,9A2 12 1C * * - ,s<L >3L LL1?L6*M L L.6 M Mc  Ub[U[5(deUnURS5(aUSSnOSURS5(aUSSnO7URS5(aUSSnOURS5(aUSS n[5(a [ XX#5$US-n[ R RU5(Gd[RUS 35 Uc [S 5eUbUVs/sHofRPM nnOSn[ R RU5(Ga[5n[ R"US S 9HMupn U HAn UR[ R R!X5R#SS55 MC MO [%UR'55n /nU Hn[ R R!XR5R#SS5nUSL=(d URU;nX;dM]U(dMfUR)U5 UR+U5 M [-U5S:aVSR![%U55n[R/SR1SR![%U5555 [3X!US9 g[ R R9U5(aUc [S5eUR'5n U Vs1sHofRiM nnUHnURU;dM[;S5e [ R R=U5unn[3UUUUS9 gSn[%[?[@UR'555nU(a=[BRDRFRIU[K5URL5 [OUS5n [PRRS:Xa*[PRTRVS:Xa [YX\5nO [[U SS9n[]U5nSSS5 UHAnURW;dSURS US!35eU"UUUR5 MC [%[?[^UR'555n[-U5S:aUS-n[ R RU5(d S"US#35eU(a=[BRDRFRIU[K5URL5 [OUS5n [[U SS9nSSS5 UHAnURW;dS$URS US!35eU"UUUR5 MC ggs snf![4an[R7U5 UeSnAf [5S5e=fs snf![4an[R7U5 UeSnAf [5S5e=f!,(df  GN=f!,(df  N=f)%a :api_attr: Static Graph This function get parameters and optimizer information from program, and then get corresponding value from file. An exception will throw if shape or dtype of the parameters is not match. This function can also load model file saved with [ save_params, save_persistables, save_vars ]. var_list can not be None when load single model file ( filename is not None When save_params, save_persistables or save_vars is called ). Args: program(Program): The program will be loaded model_path(str): The file prefix store the program executor(Executor, optional): The executor used for initialize the parameter When startup program is not run. var_list(list|tuple, optional): The Tensor list/tuple to load single model file saved with [ save_params, save_persistables, save_vars ]. Default: None Returns: None Examples: .. code-block:: python >>> import paddle >>> import paddle.static as static >>> paddle.enable_static() >>> x = static.data(name="x", shape=[10, 10], dtype='float32') >>> y = static.nn.fc(x, 10) >>> z = static.nn.fc(y, 10) >>> place = paddle.CPUPlace() >>> exe = static.Executor(place) >>> exe.run(static.default_startup_program()) >>> prog = static.default_main_program() >>> static.save(prog, "./temp") >>> static.load(prog, "./temp") Nr}rrz.json[ not found, try to load model file saved with [ save_params, save_persistables, save_vars ]zeexecutor is required when loading model file saved with [ save_params, save_persistables, save_vars ]Ftopdown\/r zvariable file [ {} ] not used)rrr zFailed to load model file, please make sure model file is saved with the following APIs: save_params, save_persistables, save_varszevar_list is required when loading model file saved with [ save_params, save_persistables, save_vars ]z/loaded var [{}] is not in program variable listrrr rzFailed to load model file , please make sure model file is saved with the the following APIs: [ save_params, save_persistables, save_vars ]. When these API called, filename CANNOT be Nonec[5RUR5R5nUR 5nUR 5(a [ RR5nGOUR5(a [ RR5nGOUR5(av[ RRR5nURUR 55 [ RRUR!55nGO-UR#5(a[ RRR5nURUR 55 [ RR%[ R&R)5R+S5SUR-55nOt[ RRR5nURUR 55 [ RR/UR155nUR3X5 g)N:r)rrrXr _place is_cpu_placerrCPUPlaceis_cuda_pinned_placeCUDAPinnedPlace is_xpu_placer Place set_placeXPUPlace xpu_device_idis_custom_place CustomPlacedevice get_devicesplitcustom_device_id CUDAPlace gpu_device_idset)rfndarrayrxrplaces r>set_varload..set_vars N # #CHH - 8 8 : HHJ >>  KK((*E  # # % %KK//1E ^^     &&(A KK #KK(():;E       &&(A KK #KK++ ((*005a8!:L:L:NE   &&(A KK #KK))!//*;rcrc@sb  z(H=== =L[))#CR(   x ( (#CR(   z * *#CR(   w ' '#CR( }}xBB&4 77>>- . .  "##~    w   2:;(3hh(N;N!N 77== $ $!eO%'WWZ%G!EA#'' T-55dC@&H $G$5$5$78  O'77<< HH=EEdCP"d*Hchh..H.>>#**3/#**84(?#a'"%((4+@"A3::o!67 %  WW^^J ' ' { '002 8H#I8HHH8H #I 88#77%I #%''-- "; Hi %$!&  0&w/@/@/BCDN 11 LNH,F,F  !4 (A <<8 #(8(8(>(>!(C)*=AI)!h?I%i0 )vv" AFF8#45H4I K " 9QVV$%  %w'8'8':; "$x/ ww~~m,, }o[ 9 ,  KK   5 5"LNH4N4N - &!)!h?I'#A66Y&  (9-J & Ay( ) $#y<@   a  "P$J"   a  "EL ) (8' &sU>V+< V0 W$+ W)2AX X/0 W!:WW!) X3X  X X,/ X=c b[U5n[5(a6[U5up#X#-nUVs/sHoUR(dMUPM nnO'[ [ [ UR555n0nUGHn[RR5RUR5nUcSURS35eURU;dMe[R"UR55n XRn U R U R :Xd/SU R SURSU R S35eU R"U R":Xd/SU R"SURS U R"S35eUR5n U R%5n [RR'5n U R)5(a[RR+5nOU R-5(ag[RR.R15nUR3U 5 [RR5UR755n O{U R95(af[RR.R15nUR3U 5 [RR;UR=55n U R?X5 S XgR'GM /nURA5HunnUU;dMURCU5 M! [EU5S :a5[FRH"S RKS RMU555 ggs snf)a Set program parameter from state_dict An exception will throw if shape or dtype of the parameters is not match. NOTICE: This function MUST called after run start_up_program Args: program(Program): The program to be set state_dict(dict): the dict store Parameter and optimizer information Returns: None Examples: .. code-block:: python >>> import paddle >>> import paddle.static as static >>> paddle.enable_static() >>> x = static.data(name="x", shape=[10, 10], dtype='float32') >>> y = static.nn.fc(x, 10) >>> z = static.nn.fc(y, 10) >>> place = paddle.CPUPlace() >>> exe = static.Executor(place) >>> exe.run(static.default_startup_program()) >>> prog = static.default_main_program() >>> static.save(prog, "./temp") >>> program_state = static.load_program_state("./temp") >>> static.set_program_state(prog, program_state) Nz Variable [ z2 ] Not found, Please make sure run startup programzVParameter's shape does not match, the Program requires a parameter with the shape of (rfrgrzXParameter's data type does not match, the Program requires a parameter with a dtype of (z ]) has a dtype of (rrzNThis list is not set, Because of Parameter not found in program. There are: {}r)'rrr#r]rrrrrrrrrXrrr rYrZrrrr is_gpu_placer rrrrrrrritemsrrrrrrr)rC state_dictparamsoptsrrfused_para_listpararr orig_para_np new_para_npten ten_placepy_placerrunused_para_listkrs r>set_program_statersN#:.J}})'2 )7K#??#Kf^W5F5F5HIJN;;++-66tyyA# $))$V W # 99 "88H$7$7$9:L$YY/K%%):):: hiui{i{h|}88< {BWXcXiXiWjjln : %%):):: jkwk}k}j~88< {BWXcXiXiWjjln : %%'C I{{++-H--// 335''))KK$$**, I&!;;001BC''))KK$$**, I&!;;//0AB GGK *()N99 %KN  "1 N "  # #A &# q  \ c c)*  !aLs N,N,cP[[[UR555$)a Get all the persistable vars from Program. Args: var(Program): The Program to get persistable vars Returns: list: The list contains all persistable vars in the program Examples: .. code-block:: python >>> import paddle >>> import paddle.static.io as io >>> paddle.enable_static() >>> data = paddle.static.data(name="img", shape=[64, 784]) >>> w = paddle.create_parameter(shape=[784, 200], dtype='float32', name='fc_w') >>> b = paddle.create_parameter(shape=[200], dtype='float32', name='fc_b') >>> list_para = io.get_program_persistable_vars( paddle.static.default_main_program() ) )rrrr)rCs r>get_program_persistable_varsrs& ~w'8'8':; <UH8nU R+USS9nU"UUU/SS 5(dM'URU5 M: 0nUHin[,R."[ R"R15R3UR45R755UUR4'Mk UsSSS5 $[RRU5(d SUS35e[9US5n[:R<S:Xa*[:R>R@S:Xa [CX85nO [EUSS9nSSS5 [GW5nUS-n[RRU5(a0[9US5n[EUSS9nSSS5 URIW5 U$!,(df  GN=f!,(df  N=f!,(df  ND=f)a6 Load program state from local file Args: model_path(str): The file prefix store the program var_list(list|tuple, optional): The Tensor list/tuple to load saved with [ save_params, save_persistables, save_vars ]. Default: None. The var_list is only used to get name, will not be modified. Returns: state_dict(dict): the dict store Parameter and optimizer information Examples: .. code-block:: python >>> import paddle >>> import paddle.static as static >>> paddle.enable_static() >>> x = static.data(name="x", shape=[10, 10], dtype='float32') >>> y = static.nn.fc(x, 10) >>> z = static.nn.fc(y, 10) >>> place = paddle.CPUPlace() >>> exe = static.Executor(place) >>> exe.run(static.default_startup_program()) >>> prog = static.default_main_program() >>> static.save(prog, "./temp") >>> program_state = static.load_program_state("./temp") r}Nrrrrrrz7var_list can not be None when model_path is a file typeFrrrc X[U[5(d [S5eURURUR UR URURR5[RRR:XaURSS9$SSS9$)Nz"value in var_list must be variableTrW)r^r rrcrXrYrZr[r_r r`rarbr\rds r>clone_var_to_block.load_program_state..clone_var_to_blocks!#x00#$HII''))))88==?dll.B.B.O.OO !%(  " $(  r=Tc[UUUUS9 g! SnUc"UVs/sHofRPM Os snfsnOUnU(a [XW-5e[R"XW-[ 5 g=f)NrTzFailed to load model/variables `%s`, please make sure model/variables file is saved with the following APIs: save_params, save_persistables, save_vars.F)r9rXr!rrRuntimeWarning)exerr r raise_error error_strrf filenamess r>_load_vars_with_try_catch5load_program_state.._load_vars_with_try_catchsM!$ '!!)   ME$+.22TcT2% #*9+@AA i&;^Ls  A-0load_program_statersLL[))#CR(   x ( (#CR(   z * *#CR( &4 77>>- . .  "##~    z : :I "$U!C DGGLL1 " F - 5 5dC @ $$]3 "D! " I"//1J ";? 8KK((*E++&&u-C Oww~~j))&(ggmmJ&?#)#C#**+=j#+NO$)?I '''...z3? (.Z$%2 $.#8#8!)t$9$5hZu,228<%2H&%'ZZKK,,.77ALLN&"' y# "| 77>>- . . .//?@ . !4 (A <<8 #(8(8(>(>!(C)*=AI)!h?I ) "),I 8+M ww~~m$$ - &!)!h?I' # c# "D ) (' &s-D+PBPA P#+ P4 P # P14 Q)rj)ryrrz Sequence[str]r{strreturnNone)r)ryrrrrrrr) rCrrTensor | list[Tensor]rrrzUnpack[_NormalizeProgramKwargs]rr)rrrrrzUnpack[_SerializeProgramKwargs]rr)F)rCrrDboolrr) rrrrrr rz$Unpack[_SerializePersistablesKwargs]rr)rCrrr rr)rrrrrr) rrrrrrrr rz!Unpack[_SaveInferenceModelKwargs]rr)rrrr)rCrrrrr rr)rrrr)r str | Nonerr rz!Unpack[_LoadInferenceModelKwargs]rz(list[Program | list[str] | list[Tensor]])....)rr rrrProgram | Noner list[Tensor] | NonerCallable[[Tensor], bool] | Nonerrrr)rr rrrrr r rr rrrr)NNNN)rr rrrrr r rr rrrr)rv) rCrrrr|rrzUnpack[_SaveKwargs]rr)NN) rCrrrrzExecutor | NonerSequence[Tensor] | Nonerr)rCrrdict[str, npt.NDArray[Any]]rr)rCrrz list[Tensor]r)rrrr rr )b __future__rrr3loggingrrrrtypingrrrrnumpyrr paddle.baserr r r r paddle.base.executorr rpaddle.base.frameworkrrrrrpaddle.base.log_helperrpaddle.framework.io_utilsrrrrrrrrio_utilsrrr r!r"pir_ior#r$r%r&r'r(r)r*collections.abcr+r, numpy.typingnpttyping_extensionsr-r.r/r1r@rFrIrMrR__all__r7INFOr7rgrrrrrrrrr rr,r0rTrr9rZrcrrrr5r=r>rsk#  :: 8.      25.).)))&y&)I) *I**i*   gll H   0#  $    D%  %    *u u$u&u. u  up 5D$5D%5D.5D 5D 5DpI 56$56%56563 56  56 56p,=^%P K K $K &K  K 0 K  K  K \ ( (X d d!d-5d d dN&R D7D7D70D7. D7 D7N $' #14  !   /      $' #14  !   /        aHaHN$( $15  !   /    D h4 h4h4h4# h4  h4 h4V !%(, Z* Z*Z*Z*& Z*  Z* Z*z _ _ "=_  _  _ D==,:>tt6t tr=